Leveraging the symbolic aggregate approximation and vector space model method for stock market classification

This research explores the integration of Symbolic Aggregate Approximation (SAX) and the Vector Space Model (VSM) to classify stock trading signals into actionable categories: buy, hold, or sell. SAX reduces the dimensionality of financial time series data while preserving essential patterns, and VSM leverages Term Frequency-Inverse Document Frequency (TF-IDF) and cosine similarity for precise classification. The study employs historical data from 25 Indonesian stocks, analyzing trends to address financial markets’ volatile and noisy nature. Key steps include data preprocessing, symbolic transformation, and vector-based classification. Results validate the efficacy of the SAX-VSM framework in identifying significant market patterns and its scalability for broader financial applications. This work highlights the framework's potential as a computationally efficient tool for informed investment decisions in dynamic market environments.

JOSEPHINE JILLIAN KOENADI Siana Halim (Advisor 1); Prof. DR. Lin, Chen-Ju (Advisor 2); I Gede Agus Widyadana (Examination Committee 1) Universitas Kristen Petra English Digital Theses Undergraduate Thesis Skripsi/Undergraduate Thesis Skripsi No. 03022687/IND/2025; Josephine Jillian Koenadi (C13210013) VECTOR SPACES; INVESTMENT ANALYSIS

Files