With the increase in the generation and spread of textual content on social media, natural language processing (NLP) has become an important area of research for detecting underlying threats, racial abuse, violence, and implied warnings in the content. The subtlety and ambiguity of language make the development of effective models for detecting threats in text a challenging task. This task is further complicated when the threat is not explicitly conveyed. This study focuses on the task of implied threat detection using an explicitly designed machine-generated dataset with both linguistic and lexical features. We evaluated the performance of different machine learning algorithms on these features including Support Vector Machines, Logistic Regression, Naive Bayes, Decision Tree, and K-nearest neighbors. The ensembling approaches of Adaboost, Random Forest, and Gradient Boosting were also explored. Deep learning modeling was performed using Long Short-Term Memory, Deep Neural Networks (DNN), and Bidirectional Long Short-Term Memory (BiLSTM). Based on the evaluation, it was observed that classical and ensemble models overfit while working with linguistic features. The performance of these models improved when working with lexical features. The model based on logistic regression exhibited superior performance with anF₁score of 77.13%. While experimenting with deep learning models, DNN achieved anF₁score of 91.49% while the BiLSTM achieved anF₁score of 91.61% while working with lexical features. The current study provides a baseline for future research in the domain of implied threat detection.

Droughts have grown increasingly common, severe, and widespread in recent decades due to climate change, aggravating their harmful repercussions. Drought prediction is very effective for providing early warning and protecting the most susceptible areas from the dangers of drought. This study looked at the feasibility of applying Deep Neural Networks to create drought stage classification models for the Tharpakar District of Pakistan. A collection of satellite pictures of Tharpakar at various degrees of the drought were employed in this investigation. The unique dataset utilized in this study was gathered utilizing the time-lapse function of Google Earth Pro. The drought stages considered in this study are 'Before Drought,' 'Drought,' 'After Drought,' and 'No Drought.' DenseNet, ResNet, InceptionV3, Xception, and VGG19 deep learning architectures were utilized for training the models. Accuracy, Precision, Recall, F1-Score, and ROC curves were used to evaluate all models. According to the experimental results, DenseNet and ResNet were the best-performing models with an accuracy of 70%, while VGG19 was the lowest-performing model with an accuracy of 60%.