PhraseLevel-Sentiment-Analysis-toolkit

This is a Python3 Theano implemented toolkit for phrase-level sentiment analysis

Dependencey

• Numpy
• Python3
• Theano

Prerequisite

Set the file path in the Makefile:

all:
	python3 run_model.py -i YOUR_INPUT_PATH -o YOUR_OUTPUT_PATH

The input directory requires seven numpy matrices:

• A.npz: sparse matrix, shape = [m_reviews, n_pairs]
• X_prime.npy: array-like or sparse matrix, shape = [m_reviews, 2]
• G.npy: array-like or vector, shape = [n_pairs]
• X_zero.npy: array-like or sparse matrix, shape = [n_pairs, 2]
• W_a.npy: array-like or sparse matrix, shape = [n_pairs, n_pairs]
• W_b.npy: array-like or sparse matrix, shape = [n_pairs, n_pairs]
• W_s.npy: array-like or sparse matrix, shape = [n_pairs, n_pairs]

Usage

There are two ways to execute the program

• The simpliest is use the command line to run the program

python3 run_model.py -i YOUR_INPUT_PATH -o YOUR_OUTPUT_PATH

• You could also amend the file path in the Makefile and do it by using make, e.g.,

> make

Reference

Zhang, Yongfeng, et al. "Do users rate or review?: boost phrase-level sentiment labeling with review-level sentiment classification." Proceedings of the 37th international ACM SIGIR conference on Research & development in information retrieval. ACM, 2014.