Data compression and prediction using machine learning for industrial IoT

Abstract

Industrial IoT generates big data that is useful for getting insight from data analysis but storing all the data is a burden. To resolve it, we propose to compress the industrial data using neural network regression into a representative vector with lossy compression. For efficiency of the compression, we use the divide-and-conquer method such that the industrial data can be handled by the chunk size of data. Through our experiments, we verify that industrial data is represented by a function and predicted with high accuracy.