Earnings Prediction with Deep Learning

TL;DR

This paper compares LSTM and TCN deep learning models for predicting future earnings per share, demonstrating their superior accuracy over naive models and analysts using financial and stock data.

Contribution

It introduces a comparative analysis of LSTM and TCN models for earnings prediction, highlighting their improved accuracy over traditional benchmarks.

Findings

LSTMs outperform naive models by up to 30%.

TCNs achieve a 30.8% improvement over naive models.

Both models surpass analyst predictions by up to 13%.

Abstract

In the financial sector, a reliable forecast the future financial performance of a company is of great importance for investors' investment decisions. In this paper we compare long-term short-term memory (LSTM) networks to temporal convolution network (TCNs) in the prediction of future earnings per share (EPS). The experimental analysis is based on quarterly financial reporting data and daily stock market returns. For a broad sample of US firms, we find that both LSTMs outperform the naive persistent model with up to 30.0% more accurate predictions, while TCNs achieve and an improvement of 30.8%. Both types of networks are at least as accurate as analysts and exceed them by up to 12.2% (LSTM) and 13.2% (TCN).