動態P/E 行為於台灣股市投資決策之研究

Abstract

以本益比高低來評估一家公司的股價是否合理，是目前國內法人及投資理財機構常用的工具，然而因比較基礎的不同，往往也出現了不同的標準，其評估的結果往往也出現相當的落差。因此市場上在運用本益比評估公司股價行為時，往往以靜態的概念來評估股價應有之位階，時有選擇低本益比的公司，而其股價成長性卻不高；反而高本益比之公司，卻出現個股爆發性成長之特殊情況。 本研究嘗試應用類神經網路具優異學習的特性，將原本是靜態的本益比固定區間投資決策行為，以實際公司本益比(P/E)與股價波動行為之關聯性，透過倒傳遞類神經網路之訓練學習後，以掌握公司未來成長性因子，並經由台灣加權指數及產業市場本益比相對於公司本益比之影響，如果能事先獲知公司未來本益比(∆P/E)的變化，將有助於衡量公司合理股票價值，改善一般運用本益比(P/E)作為買賣交易的操作模式，以獲取最佳之投資報酬。

The price-earnings (P-E) ratio is a common used tool to evaluate a value of company by the domestic legal person and the investment and finance organization currently. However, because of more basal dissimilarity, usually there appear the different standard and the gained analyzing result is also probably different. So, on the market while making use of the price-earnings ratio to evaluate the reasonable stock price of a company, people usually evaluate the level of stock price based on the concept of the static state. Sometimes there is company that chooses the low price earnings ratio, and its stock price growth is not high. However, the company with high price-earnings ratio, there appears special condition of an explosion growth in the individual stock.

This research apply Neural Network to have the characteristic of the excellent learning. It is originally a static and fixed zone investment behavior. By the network train to learn the physical running of the company’s P-E ratio and the behavior of the stock price motion and by the effect of weighted index and the industry market P-E ratio to company’s P-E ratio, we can estimate the trend behavior of the P-E ratio and the future variety of the P-E ratio. It will contribute to measuring the reasonable stock value of a company to make" the decision mode of the static state P-E ratio" applied to the type of Neural Network to help investors to increase investment return rate (IRR) by the dynamic P-E ratio operation mode in the long-term investment strategy of stock choosing.