This research delves into the urgent requirement for innovative agricultural methodologies amid growing concerns over sustainable development and food security. By employing machine learning strategies, particularly focusing on non-parametric learning algorithms, we explore the assessment of soil suitability for agricultural use under conditions of drought stress. Through the detailed examination of varied datasets, which include parameters like soil toxicity, terrain characteristics, and quality scores, our study offers new insights into the complexities of predicting soil suitability for crops. Our findings underline the effectiveness of various machine learning models, with the decision tree approach standing out for its accuracy, despite the need for comprehensive data gathering. Moreover, the research emphasizes the promise of merging machine learning techniques with conventional practices in soil science, paving the way for novel contributions to agricultural studies and practical implementations.

Fire hazard is often mapped as a static conditional probability of fire characteristics’ occurrence. We developed a dynamic product for operational risk management to forecast the probability of occurrence of fire radiative power in the locally possible near-maximum fire intensity range. We applied standard machine learning techniques to remotely sensed data. We used a block maxima approach to sample the most extreme fire radiative power (FRP) MODIS retrievals in free-burning fuels for each fire season between 2001 and 2020 and associated weather, fuel, and topography features in northwestern south America. We used the random forest algorithm for both classification and regression, implementing the backward stepwise repression procedure. We solved the classification problem predicting the probability of occurrence of near-maximum wildfire intensity with 75% recall out-of-sample in ten annual test sets running time series cross validation, and 77% recall and 85% ROC-AUC out-of-sample in a twenty-fold cross-validation to gauge a realistic expectation of model performance in production. We solved the regression problem predicting FRP with 86% r² in-sample, but out-of-sample performance was unsatisfactory. Our model predicts well fatal and near-fatal incidents reported in Peru and Colombia out-of-sample in mountainous areas and unimodal fire regimes, the signal decays in bimodal fire regimes.

This research focused on the design and implementation of the flipped classroom approach for higher mathematics courses in medical colleges. Out of 120 students, 60 were assigned to the experimental group and 60 to the control group. In the continuous assessment, which included homework and quizzes, the average score of the experimental group was 85.5 ± 5.5, while that of the control group was 75.2 ± 8.1 (P < 0.05). For the final examination, the average score in the experimental group was 88.3 ± 6.2, compared to 78.1 ± 7.3 in the control group (P < 0.01). The participation rate of students in the experimental group was 80.5%, significantly higher than the 50.3% in the control group (P < 0.001). Regarding autonomous learning ability, the experimental group spent an average of 3.2 hours per week on self-study, compared to 1.5 hours in the control group (P < 0.005). Other potential evaluation indicators could involve the percentage of students achieving high scores (90% or above) in problem-solving tasks (25.8% in the experimental group vs. 10.3% in the control group, P < 0.05), and the improvement in retention of key concepts after one month (70.2% in the experimental group vs. 40.5% in the control group, P < 0.01). In conclusion, the flipped classroom approach holds substantial promise in elevating the learning efficacy of higher mathematics courses within medical colleges, offering valuable insights for educational innovation and improvement.