ML2017FALL

Machine Learning (2017, Fall) @ National Taiwan University

HW1 (linear regression)

PM2.5 prediction

HW2 (binary classification)

Income prediction

Using the Adult dataset from:

https://archive.ics.uci.edu/ml/datasets/adult

HW3 (CNN)

Image sentiment

HW4 (RNN, word embedding)

Tweets sentiment

Using the dataset from:

http://thinknook.com/twitter-sentiment-analysis-training-corpus-dataset-2012-09-22/

HW5 (dimension reduction, matrix factorization)

Movie rating

Using the MovieLens 100K dataset from:

https://grouplens.org/datasets/movielens/100k/

HW6 (PCA, word embedding, t-SNE, autoencoder, clustering)

There are three independent parts:

1. Implementing eigenface using SVD

2. Chinese word embedding visualization (word2vec, t-SNE)

3. Dataset separation using unsupervised learning

Final project

Chinese QA based on a chinese version of the SQuAD dataset

The model is a modified version of an implementation of Microsoft R-NET:

https://github.com/matthew-z/R-net

Another model is a naïve sliding window method