Summarizing Text for Indonesian Language by Using Latent Dirichlet Allocation and Genetic Algorithm

The number of documents progressively increases especially for the electronic one. This degrades effectivity and efficiency in managing them. Therefore, it is a must to manage the documents. Automatic text summarization is able to solve by producing text document summaries. The goal of the research is to produce a tool to summarize documents in Bahasa: Indonesian Language. It is aimed to satisfy the user’s need of relevant and consistent summaries. The algorithm is based on sentence features scoring by using Latent Dirichlet Allocation and Genetic Algorithm for determining sentence feature weights. It is evaluated by calculating summarization speed, precision, recall, F-measure, and some subjective evaluations. Extractive summaries from the original text documents can represent important information from a single document in Bahasa with faster summarization speed compared to manual process. Best F-measure value is 0,556926 (with precision of 0.53448 and recall of 0.58134) and summary ratio of 30%.