The core of data analysis lies in “letting the data speak.” To achieve this, three key concepts are essential: variance

 The core of data analysis lies in “letting the data speak.” To achieve this, three key concepts are essential: variance, correlation, and the Lagrange multiplier method. Variance indicates how widely data is spread and serves as a measure of information richness. Correlation reveals common structures between variables and is useful in methods like principal component analysis (PCA). The Lagrange multiplier method is a mathematical approach to solving optimization problems under constraints and is applied in many machine learning techniques such as PCA and SVM. By combining these concepts, it becomes possible to uncover the true nature of data.