Text analytics approaches to multichannel information summarisation on Fintech customers

Abstract

This thesis focuses on the design and implementation of a system that applies machine learning techniques to financial documents with the aim of automating their summarization. The research primarily utilized the K-Means algorithm to extract key sentences for the summary of each text document in the dataset provided by Qualco SA. Additionally, the proposed method was then applied to the Financial Narrative Summarization (FNS) 2023 dataset, where it demonstrated promising results in summarizing financial narratives. The models were evaluated using metrics such as ROUGE scores to assess their effectiveness in capturing key information from the documents.