my projects
01
Standard Naive Bayes Classifier
Using python, I implemented a Standard Naive Bayes Classifier (SNB Classifier) from scratch, and used cross-validation to analyse the test accuracy of the classifier. The data used is collected from several articles from the MEDLINE database. The model is trained on this dataset and improved by my implementation of a dataclean() function on the dataset to explore whether this will improve the test accuracy, the improvement in test accuracy was validated using cross-validation as well. The trained model can take test cases and predict either Achaea, Bacteria, Eukaryota, or Virus. The implementation is very fast and has a 98% accuracy.
02
Decision Tree Classifier using Information Gain
Using python, I have built a decision tree classifier without the use of modules. Entropy and information gain calculations are used to determine splits. The model is trained using a large dataset, "Mushroom Data", obtained from the UCL Machine Learning Repository. The data used is the Mushroom Data, this dataset comes from the UCI Machine Learning repository. The model is able to receive test data and make a prediction on whether the mushroom is edible or poisonous very quickly.
03
Interactive Shape Graphics User Interface
Using Java OOP, I have implemented an interactive shape animator GUI in which the user can create or discard of shapes, as well as change the border and fill colour of each shape. The user can also make nested shapes and determine how the shape moves.
04
Extended Four-In-A-Row Game
Using Python, I developed an extended version of a four-in-a-row game where the board can have a width of up to 8, instead of four. The user plays against the computer, however, the point system of the game is much more complex than a regular Four-In-A-Row game. The board size can be changed to preference.
