I compared two machine learning models (SVM and KNN) to discover which one suit the most to be used on my data. Using scikit-learn framework to speed up the process of developing the models but manually created TF-IDF from scratch. The most suit models for my data then is used to predict student's satisfaction toward services my college has provide.
- Preprocessing
- The concept of TF-IDF for feature extraction
- Cross-validations
- How SVM and KNN works based on their algorithm