Enhancing Business Analytics through Hybrid Summarization of Financial Reports
Tohida Rehman
TL;DR
This paper tackles the challenge of distilling long earnings transcripts into concise, factually reliable Reuters-style summaries. It introduces a hybrid extractive–abstractive framework that combines LexRank-driven input selection with fine-tuned BART and PEGASUS, alongside a parallel Longformer Encoder–Decoder pathway to handle long contexts. Evaluations on the ECTSum dataset using standard metrics and entity-level factuality measures show that long-context models (LED) deliver strong overall performance, while the hybrid approach offers competitive results under computational constraints. The findings support practical, scalable systems for efficiently extracting actionable business insights from lengthy financial texts, with emphasis on improving factual consistency in automated summaries.
Abstract
Financial reports and earnings communications contain large volumes of structured and semi structured information, making detailed manual analysis inefficient. Earnings conference calls provide valuable evidence about a firm's performance, outlook, and strategic priorities. The manual analysis of lengthy call transcripts requires substantial effort and is susceptible to interpretive bias and unintentional error. In this work, we present a hybrid summarization framework that combines extractive and abstractive techniques to produce concise and factually reliable Reuters-style summaries from the ECTSum dataset. The proposed two stage pipeline first applies the LexRank algorithm to identify salient sentences, which are subsequently summarized using fine-tuned variants of BART and PEGASUS designed for resource constrained settings. In parallel, we fine-tune a Longformer Encoder-Decoder (LED) model to directly capture long-range contextual dependencies in financial documents. Model performance is evaluated using standard automatic metrics, including ROUGE, METEOR, MoverScore, and BERTScore, along with domain-specific variants such as SciBERTScore and FinBERTScore. To assess factual accuracy, we further employ entity-level measures based on source-precision and F1-target. The results highlight complementary trade offs between approaches, long context models yield the strongest overall performance, while the hybrid framework achieves competitive results with improved factual consistency under computational constraints. These findings support the development of practical summarization systems for efficiently distilling lengthy financial texts into usable business insights.