Incorporation of human knowledge to the stock markets for improving forecasts

Mitra, Swarnava and Ordieres-Meré, Joaquín (2016). Incorporation of human knowledge to the stock markets for improving forecasts. In: "Industriales Research Meeting 2016", 20 abril 2016, Escuela Técnica Superior Ingenieros Industriales - UPM - Madrid. ISBN 978-84-16397-31-0. p. 170.

Description

Title: Incorporation of human knowledge to the stock markets for improving forecasts
Author/s:
  • Mitra, Swarnava
  • Ordieres-Meré, Joaquín
Item Type: Presentation at Congress or Conference (Poster)
Event Title: Industriales Research Meeting 2016
Event Dates: 20 abril 2016
Event Location: Escuela Técnica Superior Ingenieros Industriales - UPM - Madrid
Title of Book: Industriales Research Meeting 2016
Date: 2016
ISBN: 978-84-16397-31-0
Subjects:
Freetext Keywords: social media analytics, stock market prediction, machine learning
Faculty: E.T.S.I. Industriales (UPM)
Department: Ingeniería de Organización, Administración de Empresas y Estadística
Creative Commons Licenses: Recognition - No derivative works - Non commercial

Full text

[img]
Preview
PDF - Requires a PDF viewer, such as GSview, Xpdf or Adobe Acrobat Reader
Download (1MB) | Preview

Abstract

This thesis attempts to explore the possibility to model human behavior and how it guides financial markets. According to Behavioral Finance theory the stock market ecosystem is influenced by the decision making of the individuals trading in it. The traders are heterogeneous in nature, with each group having their own belief and expectation. This thesis tries to answer the question Can human behavior and its responses to macroeconomic events be modelled and used as an indicator to predict price directions? To answer the former question, the research has delved deep into exploring human behavior guiding financial markets. Different exogenous variables representing stock broker behavior has been explored. These variables are derived from market data of local markets like the Madrid Stock Exchange, and also from micro blogging sites and website visit statistics. A local market microstructure is guided mostly by its local players and macroeconomic events. Where as more global stock markets are more guided by global macroeconomic events. This research constructs exogenous variables which effect the small stock exchanges and bigger stock exchanges alike. In this research different data set are constructed from web search volumes, sentiment scores of Twitter posts to page visit statistics of Wikipedia articles. The exogenous time series constructed is then used as a predictor variable for different supervised and unsupervised machine learning algorithms for future price predictions. In this research different categories of machine learning algorithm were used from simple tree based ensemble learning models to SVM (support vector machine) and kernel based models to more complex Deep Learning algorithms. The implication of the research is that it will help financial managers and traders use these correlations with social sentiment indexes to predict financial markets with certain accuracies. It will also provide them with early warnings of market downturns risk and indication of crisis.

More information

Item ID: 46153
DC Identifier: http://oa.upm.es/46153/
OAI Identifier: oai:oa.upm.es:46153
Official URL: http://www.industriales.upm.es/investigacion/irm16/index.es.htm
Deposited by: Memoria Investigacion
Deposited on: 30 May 2017 07:22
Last Modified: 23 Apr 2018 07:37
  • Logo InvestigaM (UPM)
  • Logo GEOUP4
  • Logo Open Access
  • Open Access
  • Logo Sherpa/Romeo
    Check whether the anglo-saxon journal in which you have published an article allows you to also publish it under open access.
  • Logo Dulcinea
    Check whether the spanish journal in which you have published an article allows you to also publish it under open access.
  • Logo de Recolecta
  • Logo del Observatorio I+D+i UPM
  • Logo de OpenCourseWare UPM