Customer Review


Customer Service


February 17th, 2017


Zenhua Huang




How to analyze sentiment of text?

To analyze the review of customers or competitive products. We need to extract the sentiment in the reviews, which is the most challenging problems in natural language problems. We need to cut the texts into parts of speech tags and then analyze the syntactic semantic trees. The topic or aspect of the texts also need to be discovered. E.g. when we analyze reviews of the products, we would like to know which aspects the user are describing: design or battery of a cell phone and how they feel about those aspects.


Build a topic and sentiment classification model.

We use the semantic parser results of texts and build a topic model to analysis the aspects of reviews. And then train a classification model to analyze the sentiment polarity of the text. Then a triple combination. Then we use visualizalization tools to show the results .


Help to improve products and services.

Automatically collecting reviews of your products and competitive products. Based on Natural Language Processing technology, we can cluster topic of the reviews and analysis sentiment of users towards different aspects of the products. And use the results of the analysis to help us improve the quality of products or service.

E.g. we found the small discount and gifts can highly improve the satisfaction of a product.

Want to learn more?

Let’s Chat


695 Town Center Drive, Suite 1200
Costa Mesa, CA 92626
United States