This image classification project is based on the famous MNIST dataset that is about handwritten digits and their corresponding labels. Before modeling, I used some preprocessing techniques to make this dataset as row data. Firstly, I reshaped each image array to single row array and then used feature extraction method, namely, Principal Component Analysis (PCA) to reduce data dimension and speed up the training process. Finally, I used Support Vector Classifier (SVC) to train the model, and GridSearchCV offered by sklearn for hyperparameter tuning.