Premium
Stock price prediction using k ‐medoids clustering with indexing dynamic time warping
Author(s) -
Nakagawa Kei,
Imamura Mitsuyoshi,
Yoshida Kenichi
Publication year - 2019
Publication title -
electronics and communications in japan
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.131
H-Index - 13
eISSN - 1942-9541
pISSN - 1942-9533
DOI - 10.1002/ecj.12140
Subject(s) - medoid , cluster analysis , dynamic time warping , search engine indexing , computer science , data mining , stock price , time series , pattern recognition (psychology) , econometrics , artificial intelligence , series (stratigraphy) , mathematics , machine learning , paleontology , biology
Various methods to predict stock prices have been studied. In the field of empirical finance, feature values for prediction include “value” and “momentum”. In this research, we use the pattern of stock price fluctuations which has not been fully utilized in the financial market as the input feature of prediction. We extract the representative price fluctuation patterns with k ‐Medoids Clustering with Indexing Dynamic Time Warping method. This method is k ‐medoids clustering on dissimilarity matrix using IDTW which measures DTW distance between indexed time‐series. We can visualize and grasp a price fluctuation pattern effective for prediction with the proposed method. To demonstrate the advantages of the proposed method, we analyze its performance using TOPIX. Experimental results show that the proposed method is effective for predicting monthly stock price changes.