Joint Optimization Method of Dimensionality Reduction and Clustering for Big Data

In recent years, the volume of data has increased rapidly, which has created unfavorable conditions for traditional data processing and analysis techniques. Clustering, as an important technique in data mining, plays an important role in handling big data. This paper reviews the joint optimization methods of dimensionality reduction and clustering for big data. Firstly, The clustering method is introduced. And then introduces three joint optimization methods: clustering and projected clustering with adaptive neighbors, projected fuzzy c-means with robabilistic neighbors, fuzzy k-means clustering with discriminative embedding, analyzes in detail the characteristics, strengths and weaknesses of each class of algorithms, and their experimental conditions are summarized. Ultimately, future research directions for joint clustering algorithms are envisioned.