About The Project

The HTGR project aims to provide members an introduction into machine learning. The HTGR team designed and developed several machine learning models including Random Forest, SVM, And LSTM models predicting sudden driving behavior.

• LSTM models use sequential data passed through memory cells to decide which information to forget and which to save as it moves to the next cell.
• Random forest models use random subsets of data labels and random feature sets to create multiple decision trees. Each tree is branched through various conditions, which might involve comparing data set values like mean acceleration, to certain thresholds.
• SVM models use a hyperplane of data, using the confidence level of the data point to create a binary classification scenario, which with high confidence could be very useful.

Using these 3 models, a broad yet highly accurate range was cast by splitting the data up into 3 groups. The project wrapped up with higher than 90% accuracy for all 3 models. To learn more about each model in depth, we highly recommend checking out the GitHub repo to learn more about each of them through the summed up powerpoint presentations on there about Machine Learning basics and details about each type of the Models we worked on.