Aplicación de dos técnicas del análisis multivariado en el mercado de valores mexicano

This article is a supplement to the stock technical analysis and its main objective is to classify 88 companies belonging to the Mexican Stock Exchange. Using principal component analysis (PCA) and linear discriminant analysis (LDA), the input hypothesis is to group companies according to their market performance and the economic sector to they belong. The methodology consisted in collecting the volume of shares traded indicator (input variables) corresponding to 88 companies for the period January 2015 to March 2016, the input data come from Infosel financial software. After that the input data were normalized and subsequently the PCA and LDA methods were applied to obtain three groups that do not meet an importance criterion. Each group has correlation with each element that makes up, but does not maintain correlation with the elements of other groups; so that if any company belonging to one of the groups presents some tendency, the other actions of the same group also showed that same trend, but companies from other groups will not tend necessarily in the same way. The results represent a significant contribution to the creation of investment portfolios. However, the authors suggest complement this analysis with the fundamental analysis approach to study issuers and reduce investment risks.