Application for VDO’s content search based on keywords
This study aimed to develop a text-based videos’ content searching application. The developed system converts videos into audio files using Google Cloud Text to Speech in order to generate .json file. Afterwards, the file is fetched to elastic search, which has the ability to search for texts. This will not only decrease the complexity of the data, but also the search length. When the user inserts the desired keywords, the application will retrieve the data from elastic search using Postman; consists of Mocking service and Ngrok for Hit matching and score calculation, and then show the results. In this study, Analyse Token was used as a tool to analyze, split texts, and evaluate the precision of the search. The test was performed on 14 dharma videos from internet using 10 random keywords. The study has shown that the VDO’s content search based on keywords with the application has the accuracy value is 80.71. Since the search used in this study is only for normal keywords, Natural Language Processing (NLP) could be used in order to enhance the performance of the application to support in the data analysis area.