Analysis and identification of ancient glass based on SVM and K-means clustering model

This paper completes the macro and micro joint classification system by establishing a Support Vector Machine (SVM) regression prediction model. On this basis, we use Principal Component Analysis (PCA) and Entropy Weight Method (EWM) method to select and calculate the weight of five dimension reduction classification indicators. We classify the types by K-means clustering algorithm. The K-center is used to calculate the classification center and euclidean distance of unknown glass, so as to obtain the calculation method of glass in cultural relics identification. Finally, we use multiple regression fitting analysis of the relationship between the components. This study is helpful to the correct identification of the glass relics and plays an important role in the study of history.