부산대학교 도서관

When reading a document text such as news or article, people tend to read related information topic. However, when they try to use search engine for looking into it, they don't have any idea what the keywords are. Copying the whole text to the search engine's bar is not the best solution, since it will increase the searching process time but make the search engine nonoptimal. It also makes people get unrelated topic. Therefore, this research has a purpose to make an application that generates keyword from inputted text of paragraph. Despite of input the whole paragraph or text in the search bar, user can input the text in the developed application and get the best keyword based on the text. The main method that has been used in this research is a text mining. First, we perform a pre-processing which are turn the inputted text to be in lower case. The second, we remove the stop words and some characters from text. The third, we perform a tokenizing method to separate each word and store it in an array. The main step is calculating the score of each word that has been stored in array. The result of this research is three words that generated by the system from inputted paragraph text. The keywords that generated by system are the same with the method's calculation. However, this application system needs to improve more to get a better performance.