The projects of Prof. Hung-yi Lee's Machine Learning course. Mainly contains different experiments over different kinds of tasks. (course website:http://speech.ee.ntu.edu.tw/~tlkagk/courses_ML17_2.html)
- Basic Machine Learning's mathematic theories
- knowledge and implement of neuron networks
- familiar to python keras, sklearn, xgboost... libraries
- brief: predict PM2.5 density in the air
- tech : implement the mathematic details of gradient descending and adagrad
- brief: analyse behavior of bank customers
- tech : implement logistic regression and XGBoost
- result: 87% accuracy of predicting will a customer retrun the money or not.
- brief: analyse one's mood by image of one's face
- tech : implement CNN and compare the results with DNN models
- result: about 65% accuracy
- brief: analyse one's mood by what one said
- tech : implement RNN and do experient over different RNN models(LSTM,GRU...etc) and Bag Of Words model.
- result: about 83% accuracy
- brief: analyse one's favor(1-10 point) over movies by semi-labeled data
- tech : semi-supervised learning, Word Embedding technique and ensemble
- result: Mean Square Error 0.85