Graphical Representation of Textual Data Using Text Categorization System

This paper presents the graphical representation of textual data using text categorization; we had concentrated on the compact representation of the document. Text Categorization has become an important task in data mining (text mining) because of the development of electronic commerce over the internet. All organizations that have business based on internet need an effective categorization method for managing large amount of textual data which is available in various forms like sales orders, summary documents, emails, journals and memos etc. Here we have used both globalized as well as localized feature selection methods. The localized method that we have introduced has also improved the accuracy of the classifier. The classifier that we have used is K-NN that is K nearest neighbor. The K-NN is simple and is having better precision in classifying a document. Also this K-NN does not need any training resources or model to be built up and it categorizes on the fly .Therefore its cost is also less as no resources need to be trained and accuracy is also better than any other classifier.