Summarization of Investment Reports Using Pre-trained Model

TL;DR

This work tackles automatic investment-report summarization by converting monthly fund-management reports into investment reports using transformer-based models. It conducts a head-to-head comparison of extractive and abstractive approaches, finding Ab-T5-based abstractive summarization superior across most metrics and fund types. The study details dataset construction from Japanese investment and monthly reports and analyzes performance by fund type, highlighting practical implications for automating formal fund disclosures. It also outlines future directions to enhance summaries with external financial indicators such as stock prices and exchange rates.

Abstract

In this paper, we attempt to summarize monthly reports as investment reports. Fund managers have a wide range of tasks, one of which is the preparation of investment reports. In addition to preparing monthly reports on fund management, fund managers prepare management reports that summarize these monthly reports every six months or once a year. The preparation of fund reports is a labor-intensive and time-consuming task. Therefore, in this paper, we tackle investment summarization from monthly reports using transformer-based models. There are two main types of summarization methods: extractive summarization and abstractive summarization, and this study constructs both methods and examines which is more useful in summarizing investment reports.

Figures (3)

• Figure 1: Investment reports.
• Figure 2: The flow of report summarization.
• Figure 3: Labeling for extractive summarization.