Analyzing customer response

I'm new to ML. I'm taking over a Classification project which involves analyzing data for customers which returned a product and I need to determine the return reason (~10 categories). This data was captured at the counter, and could include words like: LGTM (Looks good to me) NFF (No fault found), etc.

I have a training set of 1000 records and when using Google Prediction API I get a "classificationAccuracy" value of "0.82" and 10 labels.

Questions: 1. Any recommended API to analyze this type of data?. 2. What is a good "classificationAccuracy" value?

Thank you

Topic google-prediction-api classification data-mining machine-learning

Category Data Science

1
answered at

http://scikit-learn.org/stable/tutorial/text_analytics/working_with_text_data.html

You can use the above tutorial to get acquaintance with text classification. Afterwards it should be easier to formulate nontrivial questions to move even further.

About

Geeks Mental is a community that publishes articles and tutorials about Web, Android, Data Science, new techniques and Linux security.