Towards an Optimal Decision Support System
Witold Kosiński, Grzegorz Dziczkowski, Bruno Golénia, Katarzyna Wegrzyn-Wolsk · InTech eBooks · 2010
In the goal of understanding the opinions written in natural language, an Opinion Mining knowledge was necessary to implement. For this reason, we presented in this chapter new approaches to automatically detect opinion from the text. The two classifications (group conduct and linguistic) have been proposed by us. Then, we have compared our approaches with the approach generally used in this field (the statistical classification, which is based on Naive Bayes classifiers). After carrying out tests, we can observe that we have succeeded to implement a first innovative method based on a linguistic classifier. The results obtained after this classification give us satisfaction. We can, therefore, conclude that the linguistic analysis, which is deeper, is an important research path in the field of Sentiment Analysis. The final classifier can be constructed as a FUZZy-Neural Inference System (FUZZNIS) by copying the method known in approximation of multivariant functions. The designing procedure of FUZZNIS has been presented in Section 5.2. The results concerning its application to an approximation of a benchmark function of 3 variables (23) allow us to say that by applying FUZZNIS as the final classifier an optimal decision support system can be obtained.