[3K1-OS-5a] OS-5 (1)

[3K1-OS-5a-01] Node-AI: Industry-Oriented GUI Data Analysis Tool for Time-series Data

〇Yuki Miki¹, Keisuke Kiritoshi¹, Ryosuke Tanno¹, Michihiro Naito¹, Michiaki Ogawa¹, Dan Ozawa¹, Koji Ito¹ (1. NTT Communications Corporation)

Keywords:Multivariate time series data analysis, Data analysis application, Interpretability

While machine learning has begun to be applied in industries such as manufacturing due to the development of IoT, there are several problems in the popularization of machine learning in industries. Preprocessing of time series data used in the manufacturing industry and construction of models are difficult to implement, and the necessary knowledge ranges from programming technology to knowledge of time series analysis, so it is difficult to carry out the analysis unless an expert of data analysis. In this study, we developed a GUI tool, Node-AI, to cope with the popularization of data analysis to industry and the problems specific to industry. Node-AI is a data analysis tool of time series data, and it carries out the flow necessary for the application to the industry such as preprocessing of data, construction of neural network model, learning, evaluation, and factor analysis on GUI. In addition, we conducted a data analysis seminar using time-series data analysis and Node-AI as case studies.