Knowledge discovery in corporate events by neural network rule extraction

Abstract

Over the last two decades, artificial neural networks (ANN) have been applied to solve a variety of problems such as pattern classification and function approximation. In many applications, it is desirable to extract knowledge from trained neural networks for the users to gain a better understanding of the network's solution. In this paper, we use a neural network rule extraction method to extract knowledge from 2222 dividend initiation and resumption events. We find that the positive relation between the short-term price reaction and the ratio of annualized dividend amount to stock price is primarily limited to 96 small firms with high dividend ratios. The results suggest that the degree of short-term stock price underreaction to dividend events may not be as dramatic as previously believed. The results also show that the relations between the stock price response and firm size is different across different types of firms. Thus, drawing the conclusions from the whole dividend event data may leave some important information unexamined. This study shows that neural network rule extraction method can reveal more knowledge from the data. Corporate and dividend events studiesStudies show that stock prices underreact to major corporate events. Such events include dividend initiations, resumptions, and omissions [1, 2]; earnings forecast revisions [3]; stock repurchases [4]; and stock splits [5,6]. The prices continue to drift in the same direction as the announcement date reaction. Such underreaction can last up to five years after the announcement of the event. The source of underreaction, however, remains controversial. One source of underreaction is investors' irrationality or behavioral biases. Jegadeesh and Titman show that their 1993 findings continued to hold into the 1990s [7], suggesting that underreaction in their earlier study [8] is not sample-specific. Several models have been developed [9-12] to explain that underreaction can be caused by behavioral biases. For example,in [9], investors make investment decisions as though the underline firm's earnings are either in a mean-reversal regime or a momentum regime in which earnings are trending. In fact, earnings are a random walk process. Investors' misjudgments lead to underreaction in stock prices when they believe that the first regime holds and to overreaction when they believe the second regime holds.The sample-specific result of chance is the second source of underreaction. Fama argues that the behavioral models do a good job of explaining abnormal post-events performance, as they were designed to do, but cannot produce other predictions any more accurately than market efficiency hypothesis does [13]. According to the market efficiency hypothesis, stock prices reflect all available information, which implies that prices adjust quickly to new information. In [13], Fama also argues that post-event underreaction and overreaction are sample-specific results of chance. For example, Michaely et al. find long-term stock prices underreact to dividend initiation...

Interpreting Housing Prices with a MultidisciplinaryApproach Based on Nature-Inspired Algorithms and Quantum Computing

“…The most relevant difficulties in real estate price forecasting have always been the dependence on subjective judgement to precess in order to obtain reliable real estate values. For this reason, multiple regression analysis and artificial neural networks have been considered for a long time as the most flexible techniques, able to provide reliable predictions and information on real estate values and facilitate market analysis [54][55][56][57][58][59][60][61].…”

Sailing through the COVID-19 Crisis by Using AI for Financial Market Predictions

“…Some microeconomic implications can be relevant for individual behavior, especially regarding individuals who directly or indirectly influence stock market prices. A primer in the interaction of these fields had been written by Dong and Zhou [115]. Using ANNs as their point of departure, the authors deliver empirical evidence showing the relation between stock price response and firm size, which differs across different firms.…”

Using Genetic Algorithms for Real Estate Appraisals

“…For this reason, multiple regression analysis (MRA) has been considered for long time as the most flexible technique able to provide reliable predictions and information on real estate values and market analysis. As possible alternative, mostly in economic or financial fields, artificial neural networks (ANNs) have been tested by researchers for forecasting purposes [19][20][21][22][23][24][25].…”