<![CDATA[Information has become indispensable in this modern era, especially with the existence of various social media that support information update. Twitter as one of the most active social media is used to update information belonging to short text or short stories that have some difficulty when done classification, such as ambiguous word, the word contained in the test data never appear in the data train and so on. This research was conducted to determine the effect of using feature expansion or addition of word on short text in the result of classification. Prior to classification, the first data to be tested is added to the list of pre-made words as an external source or dictionary with specified limits. This limitation aims to determine the minimum value of the most optimal limit in generating the highest accuracy in the classification process. In the process of making external sources cosine similarity process is done to find the closeness between words. The result of this research is accurate showing effect of expansion of feature expansion in classification result, 83% accuracy in classification without feature expansion and increased to 87% on feature expansion with threshold value 0.9.]]>
Copyrights © 2018
