Unsupervised Learning for Lexicon-Based Classification
This work addresses the problem of improving classification accuracy for users relying on lexicons, offering a more data-driven approach without labeled data, though it is incremental as it builds on existing lexicon-based methods.
The paper tackles the lack of justification for lexicon-based classification by deriving a probabilistic model and analyzing its expected accuracy, and it introduces an unsupervised method to learn word weights from co-occurrence statistics, achieving higher accuracy than traditional heuristics.
In lexicon-based classification, documents are assigned labels by comparing the number of words that appear from two opposed lexicons, such as positive and negative sentiment. Creating such words lists is often easier than labeling instances, and they can be debugged by non-experts if classification performance is unsatisfactory. However, there is little analysis or justification of this classification heuristic. This paper describes a set of assumptions that can be used to derive a probabilistic justification for lexicon-based classification, as well as an analysis of its expected accuracy. One key assumption behind lexicon-based classification is that all words in each lexicon are equally predictive. This is rarely true in practice, which is why lexicon-based approaches are usually outperformed by supervised classifiers that learn distinct weights on each word from labeled instances. This paper shows that it is possible to learn such weights without labeled data, by leveraging co-occurrence statistics across the lexicons. This offers the best of both worlds: light supervision in the form of lexicons, and data-driven classification with higher accuracy than traditional word-counting heuristics.