Saved in:
Bibliographic Details
Main Authors: Zhang, Huopu, Liu, Yanguang, Zhang, Miao, He, Zirui, Du, Mengnan
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.14420
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911195782447104
author Zhang, Huopu
Liu, Yanguang
Zhang, Miao
He, Zirui
Du, Mengnan
author_facet Zhang, Huopu
Liu, Yanguang
Zhang, Miao
He, Zirui
Du, Mengnan
contents Predicting earnings surprises from financial documents, such as earnings conference calls, regulatory filings, and financial news, has become increasingly important in financial economics. However, these financial documents present significant analytical challenges, typically containing over 5,000 words with substantial redundancy and industry-specific terminology that creates obstacles for language models. In this work, we propose the SAE-FiRE (Sparse Autoencoder for Financial Representation Enhancement) framework to address these limitations by extracting key information while eliminating redundancy. SAE-FiRE employs Sparse Autoencoders (SAEs) to decompose dense neural representations from large language models into interpretable sparse components, then applies statistical feature selection methods, including ANOVA F-tests and tree-based importance scoring, to identify the top-k most discriminative dimensions for classification. By systematically filtering out noise that might otherwise lead to overfitting, we enable more robust and generalizable predictions. Experimental results across three financial datasets demonstrate that SAE-FiRE significantly outperforms baseline approaches.
format Preprint
id arxiv_https___arxiv_org_abs_2505_14420
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SAE-FiRE: Enhancing Earnings Surprise Predictions Through Sparse Autoencoder Feature Selection
Zhang, Huopu
Liu, Yanguang
Zhang, Miao
He, Zirui
Du, Mengnan
Computational Finance
Computation and Language
Machine Learning
Predicting earnings surprises from financial documents, such as earnings conference calls, regulatory filings, and financial news, has become increasingly important in financial economics. However, these financial documents present significant analytical challenges, typically containing over 5,000 words with substantial redundancy and industry-specific terminology that creates obstacles for language models. In this work, we propose the SAE-FiRE (Sparse Autoencoder for Financial Representation Enhancement) framework to address these limitations by extracting key information while eliminating redundancy. SAE-FiRE employs Sparse Autoencoders (SAEs) to decompose dense neural representations from large language models into interpretable sparse components, then applies statistical feature selection methods, including ANOVA F-tests and tree-based importance scoring, to identify the top-k most discriminative dimensions for classification. By systematically filtering out noise that might otherwise lead to overfitting, we enable more robust and generalizable predictions. Experimental results across three financial datasets demonstrate that SAE-FiRE significantly outperforms baseline approaches.
title SAE-FiRE: Enhancing Earnings Surprise Predictions Through Sparse Autoencoder Feature Selection
topic Computational Finance
Computation and Language
Machine Learning
url https://arxiv.org/abs/2505.14420