Konferensartikel

Active Learning for Financial Investment Reports

Sian Gooding
Dept of Computer Science and Technology, University of Cambridge, UK

Ted Briscoe
Dept of Computer Science and Technology, University of Cambridge, UK

Ladda ner artikel

Ingår i: Proceedings of the Second Financial Narrative Processing Workshop (FNP 2019), September 30, Turku Finland

Linköping Electronic Conference Proceedings 165:4, s. 25-32

NEALT Proceedings Series 40:4, s. 25-32

Visa mer +

Publicerad: 2019-09-30

ISBN: 978-91-7929-997-2

ISSN: 1650-3686 (tryckt), 1650-3740 (online)

Abstract

Investment reports contain qualitative information from numerous sources. Due to the huge volume of online business information, it is increasingly difficult for financial analysts to track and read all relevant texts. In this paper, we develop a novel tool to assist financial analysts when writing an investment report. We perform multi-class classification on business texts to categorise them into informative investment topics. Using active learning we show that we can obtain the same F1-Score of 0.74 with 58% less data.

Nyckelord

active learning, investment report, investment, machine learning, multi-class classification

Referenser

Inga referenser tillgängliga

Citeringar i Crossref