Long-term air pollution forecasting is essential for making public policies and issuing warnings. This will reduce the impact of pollution on the environment and human health. This research focuses on achieving long-term forecasting of air pollution attributes $(PM 2.5, PM_{10}, SO_{2}$, $NO_{2}, CO$, and $O_{3})$ by providing minimal historical data as input to the model. The best-performing models in this research produced between a 1% increase in RMSE for certain pollutants to a 50% decrease in RMSE for other air pollution attributes (except $CO)$ compared to the initial research on compositional learning, while being able to forecast the trend and seasonality accurately for more than one year into the future. The training time for each pollution attribute model dropped to 4 seconds compared to 1800 seconds in the case of the compositional learning method. This paper also delves into the challenges with long-term forecasting for the Beijing air quality dataset and discusses approaches to overcome these challenges.

Games live streaming is growing rapidly as a form of entertainment. A game streamer will like to know what game to stream in order to attract huge number of viewers and followers which in turn will generate sizable income for the streamer. Using streamer’s metrics, the main goal of this research work is to design and develop a set of resources that a streamer can use to maximize the number of viewers and followers for a particular game and when to play the game. This research develops two models using machine learning techniques that can be used by game streamers to maximum the returns on investment. When both model predictions are presented as percentage, Model 1 using regression algorithms provides a MAE of 5.48 meaning the prediction has an error within 5.48% of the streamer’s total follower count. Also, Model 1 has 85.46% of its predictions’ absolute error less than or equal to 5. Similarly Model 2 with 2.53 MAE and 87.68% of its predictions’ absolute error less than or equal to 5.

An efficient and cost-effective weather forecasting approach can be used to protect humans and benefit economic growth as a result of secure forest, agriculture, and tourism industry sectors. This paper is based on the IEEE Big Data IARAI’s Weather4cast 2021 challenge dataset. The goal of this paper is to consider computational cost of predicting future weather forecast by using a CNN-LSTM based neural network model. The network utilizes an encoder-decoder architecture to predict future weather images. All the four variables are predicted using the same model providing generalization in the solution. The model is trained and tested on the Nile Region (R1) data and a significant improvement is observed for the loss against cloud mask and rainfall feature prediction in comparison with CNNGRU deep learning model. Two models – shallow and deep models are compared and the results in terms of MSE values for the shallow model (which is computationally cost effective) is not too far from the deep model.

The impact of the pandemic on the airline industry has been severe. Various factors such as lockdowns, travel bans, travel restrictions and passenger footfall led to changes in the airfare. This is not limited to a few years of the pandemic as there is a possibility of a similar situation recurring in the future. To address this situation and to assess future possibilities, this paper is an attempt to apply multi-agent simulation and modeling on airfare in pandemic conditions. The objective of this paper is to develop a multi-agent model for airfare during the pandemic. We also ran simulation on the developed model based on the pandemic information available from news articles. The proposed multi-agent model has long-term utility and can be used by the airline industry, the travelers, the governments, academia, and research organizations.

Major air pollutants, especially fine particles PM2.5, are generally associated with adverse health effects, including cardiac and respiratory morbidity. The aim of this paper is to find the best combination of machine learning techniques to forecast the Air Quality Index (AQI) using the Beijing air quality datasets. The dataset consists [among other] of six air pollutant attributes - PM2.5, PM10, SO2, NO2, CO and O3that are considered important factors in calculating the Air Quality Index. Our initial results showed that Linear Regression model is not adequate in predicting and forecasting the air pollutants. Random Forest and Random Committee models performed better in terms of MAE and RMSE values compared to Linear Regression. Furthermore, it was noticed that Random Forest performs better in terms of accuracy for certain features but not all while Random Committee performs better in other set of features. This shows that using a "single" machine learning approach to predict or forecast the entire features set may not give the best accuracy. So, as a result, a modified compositional learning model with disentanglement using optimized hyperparameters and search space was designed. The results of this novel network show a marked improvement (3.34% to 78%) in terms of MAE and RMSE values when compared to Random Forest and Random Committee.

In order to provide context-aware services such as health monitoring and customized energy consumption, smart environment designers need to design robust systems for recognizing the Activities of Daily living (ADL). Once these activities are recognized, the data collected can be used for prediction. For example, energy consumption and other characteristics in the home can be predicted. This is possible if human activity in a smart home can be forecasted. The aim of this research work is to "find the best machine learning algorithm to predict human activities and to use hyperparameters tuning through performance optimization to improve the accuracy of the algorithm" The results of our initial experiments using default hyper-parameters show that Random Forest, compared to four other algorithms (MLP, SVM, Naïve Bayes, and Decision Tree), has the best accuracy at 65.32% for all features and 62.54% for reduced number of features. Tuning four Random Forest hyperparameters through optimization increases the accuracy to 97.9777% for all features and 98.287% for reduced features respectively. Using the optimized Random Forest hyperparameters, 20,000 data points are forecasted with MAE of 0.0098 compared to 0.0445 for Support Vector Machine (SVM).