hameedalmoqry's Repositories
hameedalmoqry/--Lie-Deception-Detection-
Source of the dataset :Miami University Deception Detection Database This dataset containing 320 videos of Black and White targets, female and male, telling truths and lies. Eighty (20 Black female, 20 Black male, 20 White female, and 20 White male) targets were recorded speaking honestly and dishonestly about their social relationships I have already added the features of the eye to the dataset by analysing this features from the videos using TrackEye : Real-Time Tracking Of Human Eyes ,version 2 no.of columns is 20 and no. of rows is 80 as our dataset is too small , obviously we get overfittng model I used different classification algorithm (Random Forest,Gradient Boosting,Decision Tree,AdaBoost,Logistic Regression) Random forest model was better than others with high test accuracy of 0.875