Classifying Technical Indicators Using K-Medoid Clustering

Abstract

The author uses the K-medoids method to classify technical indicators by implementing an empirical experiment on 18 different indicators. Analyzing 30 stocks, the author calculates the average distances between the indicators among these stocks using C# programming language. The K-medoids method was executed using R statistical package. The results show that the indicators were distributed among three different groups: the first group holds six indicators, the second group holds nine indicators, and the third group holds three indicators.