Simple categorization matches a document to the best fit of a small set of categories (called simple categories).
When you use simple categorization, a document is assigned to exactly one simple category. This is different to conceptual categorization, where documents might belong to zero, or multiple categories.
The simple categorization algorithm is optimized for categorizing short sections of text. In most cases, you should use ordinary categorization. However, if you are categorizing small documents, such as user comments, then simple categorization might be your best option.
Simple categorization matches a document or piece of content against your set of simple categories. For each category, it calculates a confidence score, according to how well the document matches. The document is assigned to the simple category that it matches with the highest confidence score.
You also set a default category. If a document does not have a good match to any of the simple categories, it is assigned to the default category.
You can use simple categorization for Sentiment Analysis. In this case, you set up three categories for positive, neutral, and negative sentiment. Categorization then assigns each document to one of the sentiment categories.
Simple categories are most useful for categorizing small documents. Here are a few possible uses:
categorize comments in a Web site help forum (for example, helpful, not helpful, neutral)
categorize feedback on an e-commerce Web site (for example, good service, bad service)
conduct sentiment analysis on tweets about a big news event (positive, negative, neutral)
|