Please enable JavaScript.

Coggle requires JavaScript to display documents.

AUTOMATIC NEWS ARTICLES CLASSIFICATION IN INDONESIAN LANGUAGE BY USING…

- - - - Classification
        
        Classification process
        
        Preprocessing
        
        Stopword
        
        Eliminating terms that frequently appears and are not useful in information rediscovery with the purpose of filter words that are found frequently and have low value and are not useful in information retrieval
        
        Eimination,
        
        Words weighting
        
        Every words that have been processed through preprocessing document phase then its weightings is computed to generate words that can represent a category.
        
        Parsing,
        
        Eliminates frequently found terms that are not useful in information rediscovery
        
        Case folding
        
        Conversion process of all letters in documents into small letters
        
        To unify words, eliminate noise, and lessen the vocabulary volume
        
        Stemming,
        
        Process of cutting or eliminating affixes in a word (prefixes, suffixes, infixes, and
        confixes)
        
        Naive Bayes Classifier Classification
        
        Words Weighting
        
        Every words that have been processed through preprocessing document phase then its weightings is computed to generate words that can represent a category
        
        Naive Bayes
        
        More optimum than other representational classifiers [6] such as K-means, Clustering and Cosine Similarity.
        
        This research uses 250 news articles divided into 5
        categories as learning documents
        
        Variants
        
        Supervised
        
        Document classification that has a learning method with training document in the form of learning document
        
        Unsupervised
        
        Method applied independently without training nor teaching used to analyze structure and inter data relationships
        
        Phase
        
        Learning phase
        
        Executed first to develop a learning document as a guideline for each category
        
        Learning phase has a module that resembles the classification phase. The only difference is that learning phase does not execute a classification module, but only generate documents consisting of words to characterize a category
        
        Classification phase
        
        Process
        
        Input acceptance in the form of news document texts which
        category is not known before.
        
        Process of document reprocessing, initiated by reading sentences from document texts
        
        Words, after going through preprocessing documents phases,
        are then ranked with weighting process.
        
        Process conducted a calculation based on the Naive Bayes Classifier.
      - The position of words in this method plays an important role in classification process, thus stemming method is needed (affixes elimination). Stemming is used to find the basic form of words with affixes
      - Other factors affecting the words position is learning document because the method that will be used in the application construction is a part of supervised document, a classification method using the training phase