Research on Big Data Automatic Mining Method of High-dimensional Text Based on Machine Learning

How to obtain effective and valuable topic information from massive texts has always been a hot topic in text mining. This paper attempts to use the improved algorithm to model massive high-dimensional text data. On the premise of ensuring that the effect reduction of the model is controlled within the effective range, it greatly reduces the time required for the model to run, thus improving the problem of low computational efficiency of the original model when dealing with massive data. In order to solve the problem of improving the running efficiency of LSA algorithm, this paper puts forward an improved method. In addition, this paper proves through experiments that in a big data environment, the improved method greatly reduces the time required to run the LSA method without losing the accuracy of model classification, significantly improves the efficiency of modeling, and provides a theoretical reference for the subsequent automatic mining of high-dimensional text big data.