Sports Scoring Analytics

Scoring Analytics is important in sports especially to determine the strategy of the player in competing with their opponents. Using predictive analytics, a player could use a correct strategy so that he or she can improve his/her winning rate againts any opponent. The study will emphasis on in-game scoring rather that pre-game prediction which would contribute more meaningful analytics.

Space Weather

Space weather research field is a field that studies the effects of the Sun's high-energy light and electrically charged particles on Earth and other planets. It also investigates the origins and dynamics of space phenomena such as solar flares, coronal mass ejections, geomagnetic storms and auroras. This field uses a fleet of solar observatories and satellites to observe and measure the Sun and our space.

Near-real Time Monitoring of Space Pi2 Pulsations

Pi2 pulsations are irregular magnetic fluctuations with periods of 40-150 seconds that occur in the Earth's magnetosphere. They are associated with substorm onsets and pulsed reconnection in the magnetotail. Studying Pi2 pulsations can help us understand the dynamics and coupling of the magnetosphere-ionosphere system. Therefore, this study aims to develop a machine learning predictive model for automatic identification of Pi2 pulsations to enable detections with maximal reliability and precision. 

Predictive Model of Earthquakes Based on Electromagnetic Anomalies

This study uses automated machine learning (AutoML) to develop earthquake (EQ) prediction models based on geomagnetic anomalies. The study extracts features from 50 years of geomagnetic field data using wavelet scattering transform. The study applies AutoML to select and tune the best classification algorithm among five candidates. The study finds that neural network is the best algorithm with 83.29% accuracy.

Data Analytics of Agriculture Commodities Market Price in Malaysia

Agriculture can be seen as important by being the backbone of national economy and sources of food for human. This project studied the agriculture commodities market price in Malaysia focusing on three categories namely fruit category, poultry category and vegetable category. The types of commodities chosen are berangan banana and highlands tomato for fruits, chicken and standard chicken for poultry, and choy sum and green chilli for vegetables. The selected agriculture commodities data contained information on the date, 25 places in Malaysia, three types of prices, six types of commodities, grade, unit and average price which total up to 144 887 data collected from 1 January 2018 to 5 April 2022. 12 datasets were selected from agriculture commodities market price data at the retail level for price prediction using ARIMA model.

Prediction of Steel Price Index using Machine Learning Approach

Accurate prediction of the cost material planning phase is a crucial part of a construction project’s success. One of the most significant contributors to deviations from the initial estimated cost in construction projects is material price fluctuation. Even though many researchers have highlighted the relationship between economic conditions and construction costs, an accurate prediction of the impact of this relationship to the prices of steel has yet to be achieved. This project uses four different algorithms which are PSO-SVR and ARIMAX. Here, the PSO-SVR algorithm utilizes three different types of kernel function which is linear, polynomial and radial basis function (RBF). The comparison between these algorithms proved that RBF PSO-SVR model gives the lowest value of RMSE which is USD 60.76, respectively