Data-driven outlook: Machine learning models in forecasting stock market indices

Date of Publication


Document Type

Bachelor's Thesis

Degree Name

Bachelor of Science in Management of Financial Institutions

Subject Categories

Finance and Financial Management


Ramon V. Del Rosario College of Business


Financial Management Department

Thesis Adviser

Patricia Benito

Defense Panel Chair

Tyrone Chan Pao

Defense Panel Member

Kashmirr Camacho
Ferdinand Basallo
Junette Perez
Liberty Patiu


In this research, both binary classification model and feed forward neural networks model were used in classifying whether the Philippine Stock Exchange (PSEi) closing level the next day will be higher or lower than the previous days. The researcher examined if these two machines learning models performed well in terms of (1) precision, (2) recall, (3) f1 score and (4) accuracy. The researcher focused on limiting the sample to five (n=5) because of limited research effort and time. The sample included daily stock levels of the Philippines, Japan, New Zealand, Australia, and South Korea. The other four stock markets were randomly selected out of the stock markets that closes earlier than the Philippines. The researcher used closing stock indices for years 2012-2016 of the Philippines, Japan, New Zealand, Australia, and South Korea because this is the period of post 2008 global financial crisis. The research used an open source tool for implementation of machine learning and for numerical computation using data flow graphs called Tensor Flow. The researcher obtained the data from Investing.com website, which are publicly available and free of charge. The results showed that between the two models, the feedforward neural networks model is more effective since it performed well in (a) precision (b) recall (c) f1 score and (d) accuracy. On the other hand, the binary classification model is still effective but not as much as feedforward neural networks model.

Abstract Format






Accession Number


Shelf Location

Archives, The Learning Commons, 12F, Henry Sy Sr. Hall

Physical Description

1 volume (various foliations) ; 28 cm. ; 1 computer disc ; 4 3/4 in.


Stock price forecasting--Data processing; Stock price indexed--Data processing; Stock exchanges

This document is currently not available here.