Characterising those delays proved difficult. Robust cross-correlation analysis showed only statistically detectable but weak associations between the storage-tank temperature and combustion-related operational inputs. Daily lag estimates displayed broad and heterogeneous distributions, meaning the boiler could not be described by a single deterministic response time. This thermal heterogeneity is precisely what complicates operational warning systems, since the plant’s response to combustion changes arrives at unpredictable intervals.

When the team benchmarked competing models, increasing complexity did not necessarily improve performance. A Linear autoregressive (AR) model achieved the lowest test RMSE and highest R-squared at all three horizons. Its RMSE values were 1.3774, 2.1605, and 2.6182 degrees Celsius at 60, 120, and 180 minutes respectively, with R-squared values of 0.7809, 0.4616, and 0.2104. The steady decline in accuracy with longer horizons reflects the strong persistence of the thermal system, where recent measurements carry most of the predictive information.

The study also revealed a trade-off between pointwise accuracy and coverage of delayed thermal responses. Lag coverage increased as the forecasting horizon lengthened, but predictive accuracy progressively decreased, indicating that no single horizon is universally optimal. Operators therefore face a choice between sharper short-term predictions and broader, less precise coverage of the system’s delayed behaviour.

A retrospective out-of-band warning analysis underscored this tension. XGBoost recall fell from 0.5343 at 60 minutes to 0.1761 at 120 minutes and just 0.0538 at 180 minutes, while specificity rose from 0.9095 to 0.9546 and 0.9854. In other words, longer-horizon forecasts became increasingly conservative, flagging fewer future out-of-band states but doing so with fewer false alarms.

The authors caution that multi-horizon forecasts should be interpreted as predictive risk indicators rather than evidence of causal improvements in boiler control or operational performance. For thermally delayed biomass systems, the practical message is that simple temporal models remain highly competitive, and that warning thresholds must be tuned to each horizon’s distinct accuracy and coverage profile.

Subject of Research: Multi-horizon forecasting of thermal-storage-tank temperature in a thermally delayed biomass-fired boiler system

Article Title: Multi-horizon forecasting for decision support in a thermally delayed biomass-fired boiler system

Article References: Park, S., Yang, J., Kim, S., Hwang, K., Lee, J., & Ryu, S. (2026). Multi-horizon forecasting for decision support in a thermally delayed biomass-fired boiler system. PLOS One, 21(10), e0361121. https://doi.org/10.1371/journal.pone.0361121

Image Credits: AI Generated

DOI: 10.1371/journal.pone.0361121

Keywords: biomass boiler, thermal storage, forecasting, autoregressive model, XGBoost, machine learning, time-series analysis, thermal lag, decision support, out-of-band warning, Multi-horizon, decision

The study examined typhus, dengue, leishmaniasis, malaria and epidemic encephalitis B, abbreviated EEB. Researchers obtained province- and age-specific surveillance data for 2004–2020 from the Chinese Center for Disease Control and Prevention and national aggregate data for 2021–2023 from the National Health Commission. Consequently, spatial and age-specific analyses covered 2004–2020, whereas national trend analyses extended through 2023. Across the earlier surveillance period, incidence and mortality declined significantly, with especially large reductions for malaria, dengue and EEB. Spatial hotspots were concentrated mainly in southern provinces, although the available data did not permit comparable geographic and age-specific analysis for 2021–2023.

Malaria and EEB had the largest reductions during the pandemic-period comparison, whereas typhus and leishmaniasis changed less markedly. These findings describe reported disease burden rather than experimentally measured causal effects. The analysis used Joinpoint regression to assess annual trends, seasonal-trend decomposition using STL to examine monthly seasonal patterns, and interrupted time-series analysis to evaluate monthly changes after January 2020. Seasonal autoregressive integrated moving average, or SARIMA, models trained on pre-COVID data also estimated expected monthly trajectories for 2020–2023, allowing observed pandemic-era patterns to be compared with forecasts based on earlier data.

Despite the size of the national reductions, interpreting them as a direct consequence of the COVID-19 period would be too simplistic, the researchers said. The observed changes may reflect a combination of pandemic-era public-health measures, the pre-existing long-term decline, disease-specific control programs and possible disruption to surveillance and reporting. The analysis therefore identifies a strong temporal association between the COVID-19 period and lower reported incidence and mortality, but it does not quantify the contribution of each possible factor. This distinction is particularly important because the study combines provincial and age-specific data through 2020 with national aggregate data for the final three years.

The findings underline the need for reliable surveillance and reporting during public-health emergencies, as well as forecasting tools that can be adapted when conditions change. The study does not show that the observed declines persisted independently of pandemic-era conditions or that they were produced by any single intervention. Instead, it documents a marked reduction in reported vector-borne disease burden in China, while leaving unresolved how much of the change reflected intervention effects, continuing control programs, underlying trends or changes in the completeness of health reporting.

Subject of Research: Trends in reported vector-borne disease incidence and mortality in China, 2004–2023

Article Title: Temporal trends in reported vector-borne diseases in China before and after the onset of COVID-19 from 2004 to 2023

Article References: Lei, Z., Liu, Q., & Lei, Z. (2026). Temporal trends in reported vector-borne diseases in China before and after the onset of COVID-19 from 2004 to 2023. PLOS Neglected Tropical Diseases, 20(10), e0014745. https://doi.org/10.1371/journal.pntd.0014745

Image Credits: AI Generated

DOI: 10.1371/journal.pntd.0014745

Keywords: vector-borne diseases, China, COVID-19, surveillance, malaria, dengue, leishmaniasis, epidemic encephalitis B, time-series analysis, public health, Temporal, trends

The study, published in Frontiers in Sustainable Food Systems, treats rural prosperity as more than visitor numbers or tourism receipts. Its three outcomes capture distinct aspects of rural welfare: agricultural labor productivity, measured as value added per employed person; per-capita disposable income among rural residents; and per-capita household food expenditure. Rural tourism expenditure is defined as total tourist consumption expenditure in officially designated rural zones, including spending on accommodation, food and beverage, transportation, recreational activities and local products. It is not government spending on tourism promotion or subsidies, meaning the analysis does not directly evaluate any particular fiscal subsidy program.

The models separated outcomes associated with farming efficiency from those associated with household welfare. The agricultural productivity specification included rural tourism expenditure, total tourism expenditure, rural GDP per capita, urbanization, agricultural land, irrigated area and agricultural mechanization. The rural income model included rural tourism expenditure, total tourism expenditure, rural tourist trips, rural GDP per capita and urbanization. The food consumption model included rural tourism expenditure, total tourism expenditure, rural tourist trips, rural GDP per capita, urbanization and rural income. Thus, total tourism expenditure was represented in all three models, while rural tourist flow appeared in the income and food-consumption specifications.

After logarithmic transformation and adjustment for differences and lags, the data set provided 29 annual observations for estimation, covering 1995 through 2023. The researchers used several unit-root tests to examine stationarity, an ARDL bounds test to assess long-run relationships, and dynamic ARDL simulations to trace responses to unexpected 10% changes in rural tourism spending. The ARDL framework can accommodate variables that are stationary in levels or first differences, a requirement satisfied by the series used here. Diagnostic procedures included autocorrelation, heteroskedasticity, normality and parameter-stability checks, while the simulation approach produced counterfactual response paths for each prosperity indicator.

All three models supported a long-run relationship between rural tourism expenditure and the selected prosperity indicators. The estimated long-run elasticity for agricultural labor productivity was negative, at −0.084. In the model’s log-linear interpretation, a 1% increase in rural tourism expenditure was associated with a 0.084% reduction in agricultural labor productivity, while the estimated short-run coefficient was −0.192. Rural income showed a larger long-run elasticity of −0.461 and a short-run coefficient of −0.0909. Household food expenditure had a long-run coefficient of −0.4427 and a short-run coefficient of −0.1179. These are statistical associations estimated from national annual data, not measured causal effects or results from a randomized policy experiment.

The estimated error-correction terms were negative and statistically significant in all three models, consistent with the cointegration results. Agricultural labor productivity had the most negative adjustment coefficient, at −1.409, followed by household food expenditure at −0.2664 and rural income at −0.1971. The simulated shocks produced a similarly persistent pattern. A positive 10% shock to rural tourism spending produced an immediate and continuing negative deviation in agricultural labor productivity throughout the 20-year forecast horizon, without a return to baseline. A 10% reduction in rural tourism spending was associated with a sustained increase in labor productivity. Rural household food expenditure displayed the same broad symmetry: a positive tourism-spending shock generated a persistent negative response, whereas a negative shock generated a positive response.

The income simulation was less uniform at first, showing a modest short-term positive response to a positive shock before the effect diminished, turned negative and remained adverse over the longer horizon. A 10% negative shock to rural tourism expenditure generated a sustained positive response in rural income over the same period. The researchers also conducted Granger-causality tests, which indicated temporal ordering consistent with their framework: rural tourism expenditure preceded changes in agricultural labor productivity, rural income and household food consumption in the fitted series. This statistical result describes predictive ordering within the models and should not be interpreted as proof of a causal mechanism.

The paper’s proposed explanation is a set of reinforcing mechanisms. Tourism services may draw workers away from agriculture, especially where hospitality, transport and food services offer alternatives to farming, leaving farms with less available labor. Expansion can also place upward pressure on rural land and encourage conversion of arable or permanent-crop land into accommodation and recreational facilities. The authors further describe income leakage, in which tourism receipts leave the local economy through external ownership, imported supplies or franchise arrangements, and a reinvestment trap, where households receiving tourism income channel it into buildings, mechanization or other assets rather than immediate consumption. These mechanisms are theoretical interpretations of the model results rather than pathways directly observed in the dataset.

The findings sit uneasily beside research reporting that tourism can support farm incomes, food security, livelihood resilience, organic-farming adoption or local agricultural demand. One reason may lie in measurement: this study isolates spending in rural areas rather than relying only on broader aggregate tourism indicators that can combine very different local effects. Another is time. Dynamic simulations trace adjustment across years and may capture persistent consequences that a static association misses. The authors also argue that effects may differ with the intensity, spatial form and institutional setting of tourism. Low or moderate growth, tourism closely integrated with local farming, and stronger agricultural capital may produce different outcomes from concentrated tourism development that competes for land and labor.

Diagnostic tests broadly supported the stability and specification of the models. CUSUM and CUSUM-squared results indicated stable regression coefficients and error variance, while Breusch–Godfrey and information-matrix tests did not reject the reported null hypotheses of no autocorrelation or misspecification across the assessed components. A Zivot–Andrews test identified a significant structural break in rural tourism expenditure in 2017, but the main dependent variables did not show significant level breaks, and the authors did not include break dummies in the primary specifications. The ARDL bounds F-statistics exceeded the relevant upper bounds for all three models, supporting the estimation of long-run and short-run relationships.

Several limitations constrain interpretation. The analysis contains only 30 annual observations before differencing and lag selection, leaving 29 for estimation, and it relies on one country. Lagged variables and structural controls can reduce, but cannot eliminate, reverse causality, simultaneity and omitted-variable bias. Rural income or infrastructure may increase both tourism demand and agricultural performance, while other local changes may affect all three outcomes. The models also estimate national relationships and do not directly identify particular rural communities, forms of tourism or local tourism concentrations. The authors therefore call for instrumental-variable or quasi-experimental studies to strengthen causal identification, as well as research comparing agritourism with nature-based tourism, testing poverty and inequality outcomes and identifying nonlinear thresholds at which tourism becomes detrimental. Based on the evidence presented, the defensible conclusion is that rural tourism expenditure in China was associated with weaker agricultural and household outcomes, not that tourism is beneficial or harmful in every rural setting.

Subject of Research: Associations between rural tourism expenditure and agricultural productivity, rural income and household food consumption in China

Article Title: The effect of rural tourism on rural prosperity: evidence from China on agricultural labor productivity, incomes, and household food consumption

Article References: Jin, L., Geremew, B. A., & Abate, M. C. (2026). The effect of rural tourism on rural prosperity: evidence from China on agricultural labor productivity, incomes, and household food consumption. Frontiers in Sustainable Food Systems, 10, 1904226. https://doi.org/10.3389/fsufs.2026.1904226

Image Credits: AI Generated

DOI: 10.3389/fsufs.2026.1904226

Keywords: rural tourism, agricultural productivity, rural income, food consumption, China, tourism spending, rural development, ARDL model, time-series analysis, agricultural labor, effect, rural

The analysis combined National Poison Centre records with suicide fatality data from the Department of Statistics Malaysia. Researchers used Spearman correlation, Durbin–Watson statistics and an autoregressive integrated moving average with exogenous predictor model, ARIMAX(1,0,0), to account for longitudinal autocorrelation. The reported ARIMAX result was an AIC of -1.12, while the Durbin–Watson diagnostic was 0.61. An economic impact estimation was also formulated using Malaysia’s official Value of Statistical Life as a proxy, although the supplied abstract does not report the estimated economic value.

The clearest divergence occurred between 2019 and the 2021 pandemic peak. National Poison Centre reports of non-fatal intentional poisoning attempts fell by 29.2%, from 1,212 cases to 857, while Department of Statistics Malaysia suicide fatalities rose by 85.7%. The authors interpreted this pattern as supporting a hypothesis of method substitution toward more lethal means during the pandemic; the observational analysis does not, however, establish that such substitution caused the increase in fatalities. Age-specific analysis identified people aged 20 to 39 as the strongest age-group anchor, with a correlation of rs = 0.826.

These findings are based on an ecological and time-series design covering national aggregate data over 16 years, and the available summary does not describe individual-level exposure, intent ascertainment, completeness of fatal and non-fatal case capture, or possible changes in reporting and service use. The data also cover intentional poisoning only and therefore cannot represent all suicide methods. The study’s conclusion is consequently bounded: integrating toxicovigilance information with national psychiatric surveillance may offer a cost-effective way to monitor acute behavioral shifts, but the observed weak overall correlation means that non-fatal poisoning reports should not be treated as interchangeable with official suicide mortality statistics.

Subject of Research: National time-series assessment of non-fatal intentional poisoning as a sentinel indicator for suicide mortality trends in Malaysia, 2006–2021

Article Title: Assessing non-fatal intentional poisoning incidents as a national sentinel indicator for suicide mortality trends in Malaysia (2006 to 2021): a mechanistic and economic impact estimation

Article References: Ismail, S. A., Shamsuddin, S., Md Nor, N., Crüts, B., Gamerschlag, L., Hussain, H., Sh-Hussain, S. H., Kamaruzaman, N. A., Mohd Zaini Makhtar, M., Ramdan, M. I., Ishak, N. A., Saputra, H., Che Dom, N., & Noorhisham, N. H. (2026). Assessing non-fatal intentional poisoning incidents as a national sentinel indicator for suicide mortality trends in Malaysia (2006 to 2021): a mechanistic and economic impact estimation. BMC Public Health. https://doi.org/10.1186/s12889-026-29733-8

Image Credits: AI Generated

DOI: 10.1186/s12889-026-29733-8

Keywords: Malaysia, suicide mortality, intentional poisoning, toxicovigilance, time-series analysis, sentinel indicator, method substitution, pandemic, economic impact, Assessing, non-fatal, intentional

The researchers analyzed Ecuador’s carrying-capacity factor, or LCF, as the ratio of biocapacity to the ecological footprint. In level terms, an LCF above 1 means that ecosystems’ regenerative capacity exceeds the resource demand represented by the ecological footprint; a value below 1 means the opposite. The series averaged 1.35 over the study period, with a standard deviation of 0.29, but moved from above 1.9 in the early 1990s to close to unity in recent years. The analysis used annual national data from the Global Footprint Network, World Bank and V-Dem databases. GDP per capita was measured in constant 2015 US dollars, public-sector corruption was scored on a scale from 0 to 1, the informal economy was estimated as a share of GDP, and agricultural employment was measured as a percentage of total employment.

Because the dataset contained only 32 annual observations, the team used several econometric procedures designed to accommodate short samples and gradual changes in relationships. Fourier-augmented unit-root tests indicated that all variables were integrated of order one after first differencing, while their optimal Fourier frequencies allowed smooth changes in the series’ trajectories to be represented without requiring break dates to be specified in advance. Bayer–Hanck combined cointegration tests and a Residual Augmented Least Squares autoregressive distributed Fourier lag model, or RALS-FADL, both rejected the null hypothesis of no long-run equilibrium relationship. Fully Modified Ordinary Least Squares and Canonical Cointegration Regression were then used as complementary estimators, while Fourier–Toda–Yamamoto tests examined temporal causality and wavelet coherence assessed how relationships varied across periods and frequency scales.

The clearest result supported the Load Capacity Curve hypothesis. In both long-run models, the coefficient on log GDP per capita was negative, while the coefficient on its squared term was positive, producing the predicted U-shape. Under Fully Modified Ordinary Least Squares, the linear income coefficient was −1.17 and the quadratic coefficient was 0.07, both significant at the 1% level. Canonical Cointegration Regression produced coefficients of −0.84 and 0.05, respectively, also significant at the 1% level. The pattern indicates that income growth was associated with declining carrying capacity below a turning point and improving carrying capacity beyond it. Component estimates suggest that higher income initially increased the ecological footprint but later reduced it, while biocapacity followed the reverse pattern, supporting the interpretation of the aggregate relationship without establishing a mechanism by itself.

Public-sector corruption was negatively associated with the LCF in both estimators. Its coefficient was −0.61 under Fully Modified Ordinary Least Squares and −0.41 under Canonical Cointegration Regression, with both estimates significant at the 1% level. Decomposing the ratio showed that corruption was positively associated with the ecological footprint, while its negative association with biocapacity was not statistically significant. The analysis therefore points mainly toward greater ecological demand as the component linking corruption and weaker sustainability. The study treats this as an association rather than proof of a mechanism: corruption may weaken regulatory enforcement or accountability, but the national data do not directly measure how specific governance failures altered resource use.

The informal economy also had a negative long-run association with carrying capacity, significant at the 5% level in both estimators. The estimated coefficients were −0.03 and −0.02. Unlike the corruption result, the component analysis placed informality on both sides of the ratio: it was positively associated with the ecological footprint and negatively associated with biocapacity, with all four component estimates significant at the 1% level. This pattern is consistent with a combination of greater demand for resources and weaker regenerative capacity, although the causality tests did not establish a robust directional relationship after bootstrap correction. Informality was itself an indirect estimate generated by a two-sector dynamic general-equilibrium model in which households allocate labor between formal and informal activities, adding measurement uncertainty to the interpretation.

Agricultural employment produced the study’s most counterintuitive result. Its coefficient was positive at 0.05 in both long-run estimators and significant at the 1% level, rejecting the hypothesis that it would reduce the LCF. However, the component analysis showed negative associations with both the ecological footprint and biocapacity. Reduced ecological demand would tend to improve the ratio, whereas lower biocapacity would tend to weaken it, so the positive overall coefficient cannot be assigned to either component alone. The Fourier–Toda–Yamamoto test did identify a robust temporal relationship running from agricultural employment toward the LCF, and wavelet analysis found a positive, temporally ordered relationship mainly during 2010–2015 in frequency bands 2 and 3. These patterns do not prove that farming itself improves sustainability; they may reflect productivity gains, technology adoption or other structural changes that the aggregate model cannot isolate.

Time and frequency analyses showed that the relationships were not constant throughout the 32-year study period. Wavelet coherence identified a negative relationship between GDP per capita and the LCF mainly from the late 1990s to around 2018, with arrows consistent with income preceding the LCF. A similar pattern appeared for the quadratic income term beginning in 1993. Corruption and the LCF were significantly coherent chiefly between 2007 and 2012, but the arrows did not establish clear temporal precedence. Informality preceded a negative LCF relationship during 2000–2007, mainly in frequency bands 3 to 5, while agricultural employment preceded the positive relationship observed during 2010–2015. These period-specific findings show why a single annual coefficient can summarize several distinct phases rather than one invariant relationship.

The causality results added a further distinction between association and temporal direction. Neither the linear nor quadratic income term showed robust causality toward the LCF; significance instead ran from the LCF toward income and its quadratic term under both Wald and bootstrap procedures. Corruption was bidirectional, with significance under both procedures in each direction. Informality showed no robust causality toward the LCF, while the reverse direction failed to survive bootstrap correction. Agricultural employment showed robust causality toward the LCF, but not in the opposite direction. These tests identify precedence in a time-series model rather than definitive intervention effects, so they should not be read as proof that changing one variable would cause a specified change in another.

The findings suggest that Ecuador’s environmental trajectory cannot be attributed to income alone. The U-shaped income relationship indicates that economic development can coincide with either ecological deterioration or improved carrying capacity depending on the stage reached, while the negative associations with corruption and informality point to institutional conditions associated with how much pressure production places on natural systems. Agricultural policy presents the subtler challenge: higher agricultural employment coincided with a stronger LCF, yet both components of that ratio moved in opposing directions. The study consequently favors productivity, land and water efficiency, and environmentally efficient production rather than a simplistic interpretation of agricultural employment as either beneficial or harmful.

Several limitations constrain the findings. The analysis used a single country, a small annual sample and aggregated national data, limiting generalizability and obscuring differences among regions and industries. Public-sector corruption was an expert-based index rather than a directly observed measure, and informal output was model-estimated rather than directly recorded. The study also cannot establish that a particular institutional or productive change caused the observed shifts, although the Fourier and wavelet methods were designed to reveal smooth structural changes that conventional tests might miss. Future research could test the same relationships with regional and sector-level data and add variables measuring technology, productivity, energy efficiency and resource-use practices. Within these limits, the analysis offers a detailed national-scale account of how Ecuador’s ecological capacity, institutional quality and productive structure moved together from 1991 to 2022.

Subject of Research: Relationships among income, corruption, informality, agricultural employment and ecological carrying capacity in Ecuador from 1991 to 2022

Article Title: The role of public sector corruption, informal economy and agricultural employment on the load capacity factor: Testing the LCC hypothesis for Ecuador

Article References: Toledo, E., Conde, F., Alvarado, R., Tillaguango, B., Işık, C., & Rehman, A. (2026). The role of public sector corruption, informal economy and agricultural employment on the load capacity factor: Testing the LCC hypothesis for Ecuador. Environmental and Sustainability Indicators, 32, Article 101565. https://doi.org/10.1016/j.indic.2026.101565

Image Credits: AI Generated

DOI: 10.1016/j.indic.2026.101565

Keywords: Ecuador, ecological carrying capacity, Load Capacity Curve, economic growth, public-sector corruption, informal economy, agricultural employment, biocapacity, ecological footprint, time-series analysis, role, public

Relative humidity had an approximately U-shaped association with incidence. The pollutant exposure–response curves also differed from one another: sulfur dioxide and nitrogen dioxide had S-shaped relationships, while carbon monoxide showed a continuously rising association with incidence. All five environmental factors had statistically significant lagged effects within particular lag windows, indicating that their modeled relationships with reported cases were not confined to the same day.

Seven meteorology–pollutant pairs were evaluated for interaction. Only the mean-temperature–carbon-monoxide combination showed statistically significant synergy. Its relative excess risk due to interaction was 0.237, with a 95% confidence interval of 0.032 to 0.477, while its attributable proportion due to interaction was 0.232, with a 95% confidence interval of 0.039 to 0.39.

The other five tested pairs did not have statistically significant interactions. These model estimates indicate a joint association beyond the additive combination of the two factors in the analyzed data; they do not establish that temperature or carbon monoxide caused mumps infections or that either factor directly increased an individual’s risk.

Subject of Research: Environmental associations with reported mumps incidence in Chongqing, 2015–2024

Article Title: Interactive effects of meteorological factors and air pollutants on mumps incidence in Southwest China, 2015–2024

Article References: Yang, C., Yang, L., Wang, N., Yang, S., Tang, X., & Qi, L. (2026). Interactive effects of meteorological factors and air pollutants on mumps incidence in Southwest China, 2015–2024. BMC Infectious Diseases. https://doi.org/10.1186/s12879-026-14314-2

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14314-2

Keywords: mumps, Chongqing, air pollution, carbon monoxide, temperature, relative humidity, infectious disease, environmental epidemiology, time-series analysis, Interactive, effects, meteorological

The study, a retrospective time-series analysis spanning nearly a decade, drew on the District Health Information System, the digital backbone of Uganda’s public health surveillance. From January 2016 through 2025, the researchers compiled monthly counts of stroke cases and stroke deaths reported across the country. Because raw case counts can rise simply because more people live in Uganda each year, the team estimated monthly population figures using exponential growth interpolation and folded them into their statistical models as an offset term. That step allowed them to separate genuine increases in disease risk from the arithmetic of a growing population.

The modeling framework was Bayesian negative binomial regression, a choice suited to monthly count data that are typically overdispersed, meaning the variance exceeds the mean. The models yielded incidence rate ratios with 95 percent confidence intervals, quantifying how much the risk of stroke incidence or death changed month over month. For incidence, the ratio was 1.005 (95 percent CI: 1.004 to 1.006); for mortality, it was 1.006 (95 percent CI: 1.005 to 1.007). Both figures carry p-values below 0.001. The numbers look small, but their effect compounds relentlessly: a monthly increase of half a percent, sustained for years, translates into a substantially larger burden over a decade.

To understand what was driving the rise, the researchers turned to decomposition analysis using the Kitagawa method, a classical demographic technique that partitions change in a rate into contributions from population composition and from the rates themselves. Applied to the model estimates, the decomposition delivered an unambiguous verdict: the epidemiological effect was roughly twice as large as the population effect. In other words, most of the growth in stroke cases and deaths stems from rising incidence rates, not merely from more Ugandans being alive to fall ill. That distinction matters enormously for policy, because it implies that prevention, not demographic inevitability, holds the key to reversing the trend.

Before trusting the trends, the team subjected the data to rigorous diagnostic checks. Stationarity was assessed with unit root tests, and autocorrelation in the residuals was evaluated with the Ljung-Box test, standard procedures for confirming that a time series behaves in a statistically coherent way. Only after these checks did the researchers proceed to forecasting, using the exponential smoothing state-space model, known as ETS, which generates predictions with 80 percent prediction intervals that capture the plausible range of future values rather than a single point estimate.

The projections are sobering. By December 2028, monthly stroke cases are forecast to reach approximately 927, with an 80 percent prediction interval spanning 709 to 1,156. Monthly deaths are projected at roughly 126, with an interval of 94 to 160. The width of those intervals reflects genuine uncertainty in extrapolating health-system data, but even the lower bounds represent a marked escalation from current levels. If the trend holds, Uganda’s hospitals and stroke units will face a caseload in 2028 that dwarfs what they managed even a few years ago.

Stroke is already a leading cause of death and disability worldwide, and the global burden has shifted decisively toward low- and middle-income countries, where prevention infrastructure, acute care capacity, and rehabilitation services are often thinnest. Uganda exemplifies the pattern. Hypertension, diabetes, and other cardiovascular risk factors are rising across sub-Saharan Africa as diets, activity levels, and urbanization change, yet many patients never receive a diagnosis until a catastrophic vascular event occurs. The new findings give that familiar narrative a hard quantitative edge: the epidemic is not approaching, it is already underway and accelerating.

The authors are explicit about the implications. In the absence of effective prevention measures, the stroke burden will continue to climb through 2028 and likely beyond. They call for strengthened primary prevention, meaning population-level control of blood pressure, better detection and management of diabetes and atrial fibrillation, tobacco control, and dietary interventions. They also urge expanded acute stroke care capacity, including timely imaging, thrombolysis, and stroke units, which remain scarce across much of the region. And they emphasize improved surveillance, since the DHIS2 facility data that made this analysis possible still capture only part of the true disease picture, particularly among patients who never reach a health facility.

What makes the study notable beyond Uganda is its methodological clarity. By combining Bayesian count models, demographic decomposition, and formal forecasting within a single analytical pipeline, the researchers produced evidence that is both statistically defensible and directly actionable. The finding that epidemiological change outpaces population growth by a factor of two is the kind of result that can redirect national health planning: it says the crisis is behavioral and clinical, not demographic, and therefore preventable.

For a country whose health system already strains under infectious disease burdens alongside a growing tide of noncommunicable conditions, the forecast is a warning with a deadline. Every month of inaction adds roughly half a percent to stroke risk, compounding silently across a population of tens of millions. The mathematics of the trend is unforgiving, but so is the opportunity: interventions that bend the incidence curve even slightly would avert thousands of cases and deaths over the coming years. Uganda’s stroke epidemic, the data make clear, will be shaped by what happens now, not by fate.

Subject of Research: Rising stroke burden and forecasts in Uganda

Article Title: Rising stroke burden in Uganda: A national time-series analysis and forecasts, 2016–2028

Article References: Nakasendwa, S., Mutyoba, J. N., Ario, A. R., Omondi, E., Bashemera, B. B., Kakooza, F., Najjemba, L., Nabende, J. N., Guwatudde, D., & Kiragga, A. (2026). Rising stroke burden in Uganda: A national time-series analysis and forecasts, 2016–2028. PLOS One, 21(10), e0360963. https://doi.org/10.1371/journal.pone.0360963

Image Credits: AI Generated

DOI: 10.1371/journal.pone.0360963

Keywords: stroke, Uganda, time-series analysis, forecasting, noncommunicable diseases, public health surveillance, DHIS2, Bayesian regression, epidemiology, stroke prevention, health systems, sub-Saharan Africa

The study, published in Discover Artificial Intelligence, was led by Mohamed Amine Ouassil and colleagues at Hassan II University of Casablanca together with collaborators at Chouaib Doukkali University and other Moroccan institutions. Their central insight is that dropout prediction has long suffered from a hidden trade-off. Models evaluated primarily on final accuracy metrics tend to perform their best work only after a course has nearly run its course, by which point any warning arrives too late to help. The researchers argue that a genuinely useful early warning system must balance two competing goals: predictive accuracy and intervention timeliness. Their framework, called MI-CBiL, is designed to let institutions choose where to sit on that spectrum rather than forcing a single compromise.

Technically, the architecture processes two fundamentally different kinds of information in parallel streams. The first is a dynamic temporal interaction matrix, a day-by-day record of nine behavioral features such as logins, video views, forum participation, and quiz attempts, essentially a time-series portrait of how intensely a student engages with the platform. The second is a static enrollment-history vector capturing time-invariant facts about the student’s past: the number of previous courses attempted, completion rates, cumulative days spent on prior registrations, and historical log counts. Students with no prior platform history receive zero-valued counts and global mean rates, ensuring the model can still operate on newcomers.

The dynamic branch begins with a one-dimensional convolutional neural network whose filters slide along the temporal axis, detecting short-term behavioral motifs such as a localized burst of video interactions immediately followed by an assessment attempt. Convolutional activations pass through max-pooling, which retains the strongest local responses while halving the sequence length, and a dropout layer that randomly deactivates neurons during training to prevent overfitting. The pooled sequence then feeds a bidirectional long short-term memory network, which processes the trajectory both forward and backward, capturing contextual information from past and future states through its gated forget, input, and output mechanisms. The final hidden states from both directions are concatenated and fused with the static history vector before a sigmoid output layer produces the dropout probability.

What distinguishes the framework from prior CNN and LSTM hybrids is not any single neural operator but a decision-level innovation: a weighted multi-objective criterion for selecting the prediction window. For each candidate observation window k, the objective function J combines a misclassification penalty, weighted by classification accuracy, with a timeliness penalty proportional to k divided by the maximum observation horizon. A scalar weight alpha controls the trade-off: values near one prioritize accuracy, values near zero prioritize early detection. The optimal window is the one that minimizes the combined penalty. Because alpha is treated as a policy parameter, educational institutions can configure the system according to their own preferences, from rapid screening to high-precision detection.

The experimental evaluation drew on the KDD Cup 2015 dataset, a well-known benchmark containing approximately 120,542 enrollment records from 79,186 unique students across 39 courses on the XuetangX platform. Each course was trained and tested separately, allowing the identification of course-specific optimal windows that account for unique subject-matter dynamics. The full MI-CBiL model achieved 95.20 percent accuracy and an F1-score of 97.00 percent, surpassing single-stream baselines including standalone CNN, BiLSTM, and combined CNN-BiLSTM architectures. Under five-fold cross-validation, the multi-input model also exhibited markedly lower variance, with an accuracy standard deviation of 0.07 compared to 0.17 for the single-stream CBiL baseline, and paired t-tests confirmed the improvements were statistically significant across all matched folds.

The ablation results reveal how each component contributes. The static history module alone reached 91.93 percent accuracy, demonstrating that past behavior carries substantial predictive information. Adding the temporal branch lifted the full model 3.27 percentage points above the static-only configuration, indicating that current-course engagement patterns provide complementary signals that historical data cannot capture. The convolutional and recurrent components proved complementary as well: the combined CBiL architecture improved on standalone CNN by 0.32 points and standalone BiLSTM by 0.69 points, while the multi-input extension added a further 0.90 points on top.

Perhaps the most striking findings concern the two operational configurations. Under the Early Warning setting, with alpha set to 0.01, the objective function selected Day 1 as the optimal observation window for all 39 courses, achieving a weighted average accuracy of 93.11 percent. In other words, the very first day of interaction data, combined with enrollment history, already contains enough signal to identify at-risk students with over ninety percent accuracy. The researchers are careful to stress that this should not be read as proof that dropout outcomes are fully determined by first-day behavior; rather, Day 1 predictions are preliminary risk signals suited to low-cost interventions such as reminder messages, personalized check-ins, and encouragement to re-engage. Under the High Precision setting, with alpha at 0.99, the framework selected an average window of 19.24 days and reached 95.34 percent accuracy, slightly exceeding the Day 30 reference while cutting the required observation period by 35.9 percent.

A sensitivity analysis across intermediate alpha values revealed that the relationship between the weight and the chosen window is non-linear, exhibiting threshold-like behavior rather than a smooth progression. At alpha values of 0.01 and 0.50, Day 1 was selected everywhere; at 0.90 the average window grew to 1.38 days, at 0.95 to 3.79 days, at 0.97 to 9.21 days, and at 0.99 to 17.87 days. This validates the criterion as a configurable policy mechanism: institutions demanding immediate intervention can adopt an early-warning configuration, while those requiring more evidence before acting can push the decision later in exchange for modest accuracy gains. Computational efficiency analysis reinforced the practical viability: while training time rose linearly from 10.39 seconds at k=1 to 64.06 seconds at k=30, inference remained remarkably stable between 0.75 and 1.06 seconds, low enough for real-time deployment in live MOOC environments.

The authors acknowledge important limitations. The KDD Cup 2015 dataset, though a challenging and widely used benchmark, predates the massive transformation of online education after 2020, and today’s platforms incorporate more mobile learning, personalized recommendations, live synchronous activities, and shorter learning sessions. All 39 courses came from a single platform environment, so the results demonstrate within-platform robustness rather than full temporal or cross-platform generalizability. Future work will validate the framework on more recent datasets and across platforms, reconciling shared behavioral events and testing robustness across different learner populations and course formats. The study also evaluates prediction performance only, not the effectiveness of particular interventions, leaving real-world testing of the actions triggered by these warnings as a critical next step. Still, by shifting the focus from post-hoc analysis to real-time, early-stage detection, the MI-CBiL framework offers MOOC providers a computationally efficient, strategically flexible tool for keeping students engaged, and for reaching them before they quietly disappear.

Subject of Research: Artificial Intelligence

Article Title: Balancing early detection and performance in predicting students at risk of dropout on MOOCs using time-series analysis and a hybrid deep learning model

Article References: Ouassil, M. A., Jebbari, M., Errami, M., Rachidi, R., Hamida, S., Cherradi, B., & Raihani, A. (2026). Balancing early detection and performance in predicting students at risk of dropout on MOOCs using time-series analysis and a hybrid deep learning model. Discover Artificial Intelligence, 6(1), Article 1528. https://doi.org/10.1007/s44163-026-02480-w

Image Credits: AI Generated

DOI: 10.1007/s44163-026-02480-w

Keywords: MOOC dropout prediction, deep learning, CNN, BiLSTM, time-series analysis, early warning systems, educational technology, KDD Cup 2015, Balancing, early, detection, performance

The research, conducted by Yun Yu, Jianfeng Yang, and Benyan Xu of the China Geological Survey Development and Research Center, together with Haojie Liu of the Office of Geological Survey Planning at the China Geological Survey, examines annual data on lithium exploration from 2015 through 2024. Rather than treating exploration investment as a single undifferentiated flow of money, the team disaggregates total exploration expenditure into three distinct components: central government funding, provincial government funding, and enterprise-side spending. This decomposition matters because each channel operates under different institutional logic, different budgeting cycles, and different incentives, and the study shows that these differences translate into visibly different responses to changing lithium price conditions.

The analytical approach is deliberately descriptive. The authors track how the relative contributions of the fiscal and enterprise channels to total exploration expenditure shift over time, and they relate those shifts to two distinct temporal representations of lithium price conditions. The first is the lagged price level, meaning the prices observed in prior periods, which capture the accumulated price environment that budget makers and corporate planners responded to when their decisions were actually made. The second is the annual price change, which captures the momentum and direction of price movements within a given year. Separating these two representations turns out to be essential, because the study finds that they are associated with opposite patterns in financing composition.

The headline finding is a divergence. Higher lagged lithium prices are associated with increased public exploration expenditure and, at the same time, a decline in the enterprise financing share of total exploration spending. The mechanism the authors identify is one of relative speed: public appropriations increase more rapidly than enterprise-side spending when prior prices have been high, so even if corporate money is not shrinking in absolute terms, the public sector’s slice of the total grows. This is a subtle but consequential result. It means that a boom in lithium prices does not simply attract more private exploration capital in this institutional setting; it also triggers a fiscal response strong enough to crowd the enterprise share downward as a proportion of the whole.

By contrast, when the authors look at annual lithium price increases rather than lagged price levels, the association flips. Years in which lithium prices rose are associated with a higher enterprise financing share of exploration expenditure. In other words, the short-run direction of prices appears to move in step with the relative weight of corporate funding, while the accumulated price environment of previous periods moves in step with the relative weight of public funding. The two channels are not merely different in magnitude; they respond to different aspects of the price signal, and they respond in opposite directions when measured as shares of total spending.

Component-level analysis reinforces this interpretation. The descriptive patterns are consistent with differences in how public appropriations and enterprise-side spending evolved over the study period. Public funding, flowing through multi-level planning and budgeting processes, appears to adjust on a timetable and with a momentum that differs fundamentally from the way corporate exploration budgets are set. Central and provincial fiscal channels each carry their own planning rhythms, and the study’s disaggregation makes it possible to see that the aggregate trajectory of exploration spending conceals these channel-specific dynamics rather than revealing them.

Why does this matter beyond the statistics? Exploration is the riskiest and earliest stage of the mineral supply chain, the phase where geological uncertainty is highest and returns are most distant. In many market economies, this stage is dominated by junior mining companies financed by equity markets that are acutely sensitive to commodity price signals. In a setting characterized by multi-level state planning and significant public geological survey programs, however, the fiscal channel can play a leading role, and the study documents that this channel behaves differently from enterprise funding under exactly the price conditions that matter most for supply security. If analysts and policymakers rely solely on aggregate exploration expenditure, they may misread the underlying dynamics, mistaking a public-sector surge for a market-driven boom, or overlooking a relative retreat of private capital that is masked by rising totals.

The implications extend to how critical mineral policy is designed and evaluated. Governments around the world are currently weighing how much public money to direct toward securing supplies of battery metals, and the study offers a rare empirical window into how fiscal and enterprise channels interact in a system where both are substantial. The finding that public appropriations respond more rapidly than enterprise spending to high lagged prices suggests that fiscal mechanisms in this setting may function as a stabilizing or anticipatory force, stepping in when price conditions signal scarcity, while enterprise funding’s association with contemporaneous price increases suggests a more reactive posture. Whether similar divergences exist in other national contexts remains an open question, but the study provides a template for asking it: disaggregate the financing, distinguish lagged levels from annual changes, and compare shares rather than totals.

The authors are careful and explicit about the limits of what their evidence can support. The annual time series is short, spanning only a decade, and the associated data limitations are real. They emphasize that the findings should be interpreted as suggestive descriptive empirical associations rather than as evidence of causal or generalizable behavioral relationships. The patterns are specific to a single national setting and to the lithium exploration sector, and they should not be read as generalizable beyond that context without further evidence. This caution is worth underscoring in an era when critical minerals research is often swept into sweeping narratives; here, the contribution is a disciplined documentation of patterns within a well-defined institutional environment, not a universal law of mineral finance.

Even with those caveats, the study’s core message lands with force: financing composition matters, and aggregate expenditure is an incomplete lens. As the world’s appetite for lithium continues to grow, understanding who actually pays for the search for new deposits, and how different kinds of funders respond to the same price signals, will be essential for anyone trying to forecast supply, design policy, or invest in the energy transition. This research shows that in at least one major lithium exploration system, the fiscal and enterprise channels are not two parts of a single machine but two machines running on different clocks, accelerating and decelerating in response to different facets of the market. Recognizing that divergence is the first step toward understanding it, and toward building the kind of evidence base that critical mineral strategy will ultimately require.

Subject of Research: How fiscal and enterprise financing channels for lithium exploration respond differently to lithium price changes

Article Title: Divergent financing responses in lithium exploration: descriptive time-series evidence from fiscal and enterprise channels (2015–2024)

Article References: yu, Y., yang, J., liu, H., & xu, B. (2026). Divergent financing responses in lithium exploration: descriptive time-series evidence from fiscal and enterprise channels (2015–2024). Humanities and Social Sciences Communications. https://doi.org/10.1057/s41599-026-09378-z

Image Credits: AI Generated

DOI: 10.1057/s41599-026-09378-z

Keywords: lithium, critical minerals, exploration financing, commodity prices, China Geological Survey, public expenditure, enterprise investment, energy transition, mineral supply chains, time-series analysis, fiscal policy, battery metals

The research team, led by Luísa Maria de Souza Viana of the Universidade Estadual do Norte Fluminense Darcy Ribeiro, compared multiple linear regression, autoregressive moving average (ARMA), autoregressive integrated moving average (ARIMA), and seasonal ARIMA (SARIMA) models against feedforward multilayer perceptron (MLP) neural networks. Their goal was to reconstruct monthly discharge at the Ituberá gauging station, located on the Cachoeira Grande River approximately 21.2 kilometers upstream from the estuary mouth. The station drains a gauged area of 263 square kilometers within a basin of roughly 473.68 square kilometers, situated in a tropical humid climate with mean annual precipitation of about 2,095 millimeters and no marked dry season.

The timing of the study gave it an unusual natural experiment. Northeast Brazil has endured a severe and persistent drought beginning in 2012, characterized by widespread precipitation deficits and sustained hydrological stress. The researchers designed two temporal configurations: T1, which calibrated models on data through 2005 and validated on the subsequent record, and T2, which extended calibration through 2011 and focused validation on the post-2012 drought period. This allowed them to test whether models trained on longer, more recent records would generalize better under hydroclimatically contrasting conditions—a question of growing urgency as climate extremes intensify worldwide.

The methodological rigor of the study stands out. Rather than relying on a single validation split, the team employed rolling-origin cross-validation, an expanding-window procedure that preserved temporal ordering and evaluated candidate models at multiple forecast origins within the calibration period. Candidate MLP configurations varied in predictor set, hidden units, and weight-decay regularization, with three fixed-seed random initializations per candidate per fold. The final selected networks were fitted twenty times and averaged into ensemble predictions, reducing dependence on any single initialization. A seasonal-naive benchmark—simply repeating the discharge observed in the same calendar month of the previous year—provided a minimum-performance reference that all fitted models had to beat.

The headline finding is striking in its simplicity: no model family was uniformly superior. Under T1, the univariate MLP using three discharge lags achieved the lowest cross-validation errors of all model families, with a root mean squared error of 1.713 cubic meters per second. Under T2, however, the SARIMA model dominated the selection stage with a cross-validation RMSE of 1.309 cubic meters per second—roughly 30 percent lower than its ARMA and ARIMA competitors. Yet when these same models faced independent validation on held-out data, the rankings reversed. The T2 SARIMA’s selection-stage advantage evaporated, producing validation errors similar to simpler models, while the univariate MLP edged ahead of the stochastic models under T2 but fell behind them under T1.

Statistical testing drove the point home. Paired Diebold–Mariano tests, which compare predictive accuracy using absolute-error loss differentials over matched dates, detected no significant differences between the univariate MLP and any of the stochastic models under either temporal configuration after correction for multiple testing. Even more revealing were the predictor-matched comparisons, in which linear regression and MLP models were given identical input variables. Under T1 with discharge-plus-precipitation predictors, linear regression significantly outperformed the MLP—the only statistically significant result among the matched contrasts. The nonlinear transformation, in other words, added flexibility without adding accuracy.

The study also delivered a sobering hydroclimatic signal. Mean monthly discharge dropped from 5.52 cubic meters per second before 2012 to 4.71 cubic meters per second afterward—a reduction of approximately 14.8 percent. Thirteen post-2012 months exhibited discharge below the 10th percentile of the pre-2012 distribution, and both Welch and Wilcoxon tests confirmed the reduction was significant after adjustment. Yet the Brown–Forsythe test found no significant change in variance, the monthly climatology retained its general shape, and formal change-point tests pointed to candidate breakpoints in 1981 or 1985 rather than 2012. The post-2012 period, the authors conclude, represents a hydroclimatically contrasting interval of reduced discharge magnitude rather than a formally demonstrated regime shift.

Perhaps the most technically instructive results concern error propagation. One-step-ahead predictions, which used observed antecedent discharge as input, transferred reasonably well to held-out data. But completely recursive trajectories—where model-generated estimates were fed back into the predictor history—accumulated errors rapidly, with several configurations producing negative Nash–Sutcliffe efficiency values. The univariate MLP, for instance, produced recursive RMSE values of 3.810 and 3.401 cubic meters per second under T1 and T2 respectively, far worse than its one-step performance. Multi-horizon predictions at 1, 3, 6, 12, and 24 months showed a general decline in skill as distance from the last observation increased. The lesson: satisfactory one-step generalization cannot be extrapolated to long uninterrupted reconstructions.

The team quantified predictive uncertainty using a nonparametric moving-block residual bootstrap restricted entirely to calibration information. Pseudo-out-of-sample residuals were organized into overlapping 12-month blocks to preserve short-term dependence and annual structure, then resampled with replacement across 2,000 realizations. The resulting 95% predictive intervals achieved empirical coverage between 96.2 and 97.4 percent—slightly conservative but broadly stable across configurations. These intervals, the authors emphasize, remain conditional on the selected models, predictors, and calibration procedure; they do not incorporate uncertainty from model selection, alternative architectures, or structural hydrological choices.

For the Serinhaém estuary, the implications are conceptual but consequential. River discharge controls estuarine circulation, salinity gradients, stratification, residence time, and the delivery of nutrients and contaminants. Persistent reductions in freshwater inflow may favor landward salinity penetration and alter the time available for sedimentation and biogeochemical processing—changes relevant to an estuary within the protected Pratigi Environmental Protection Area. But the authors are careful: the approximately 15 percent reduction cannot be translated into specific salinity displacement or residence-time changes without calibrated hydrodynamic modeling, and the reconstructed series serves as an upstream boundary indicator rather than a complete estuarine water budget. The broader message for the field is equally clear: in data-scarce tropical coastal basins, rigorous temporal validation, predictor-matched benchmarking, and explicit uncertainty characterization matter more than assuming that more complex nonlinear models are inherently better.

Subject of Research: Environmental Science

Article Title: Modeling freshwater inflow to a tropical estuary using stochastic and neural-network approaches under contrasting hydroclimatic conditions in Brazil

Article References: Viana, L. M. D. S., Tostes, E. C. L., Constantino, W. D., Luze, F. H. R., Salomão, M. S. M. D. B., de Jesus, T. B., & de Carvalho, C. E. V. (2026). Modeling freshwater inflow to a tropical estuary using stochastic and neural-network approaches under contrasting hydroclimatic conditions in Brazil. Environmental Monitoring and Assessment, 198(11), Article 1211. https://doi.org/10.1007/s10661-026-16034-1

Image Credits: AI Generated

DOI: 10.1007/s10661-026-16034-1

Keywords: hydrology, neural networks, time-series modeling, estuary, Brazil, drought, machine learning, streamflow reconstruction, Modeling, freshwater, inflow, tropical