RESULTS AND DISCUSSION Results
In this study the prediction of stock returns using financial ratios based on historical cost compared to modified expense (accounting for inflation) was studied with neural network approach. Results Descriptive statistics for the variables showed that the average return on equity in the companies surveyed is 29.168. As observed ratios adjusted based on the general price index is lower than the ratios have been modified. For example, the average return on average assets and return on assets adjusted to 0.112 based on the general price index that reflects the 0.094 is that the denominator of this ratio are assets based on the general price index adjustment and have been updated this has led to reduced yields and adjusted. The results predict stock returns based on financial ratios based on historical cost through least squares regression shows between return on equity and return on assets with fixed asset turnover ratio is negative and significant. Between stock returns and the current ratio, asset turnover ratio and the ratio of net income to sales, there is a significant positive relationship. The results predict stock returns based on financial ratios adjusted for general price index by least squares regression showed that the current ratio-adjusted stock returns, asset turnover ratio, the ratio of net income to sales and return on assets and a significant positive relationship exists. The negative correlation between stock returns and fixed asset turnover ratio is significant. The results predict stock returns based on financial ratios based on historical cost through neural network showed the network has good performance in predicting stock returns variable and fixed asset turnover ratio, asset turnover ratio, return on assets, current ratio and net profit margins, respectively, are of the greatest importance and impact. The results predict stock returns based on financial ratios adjusted through neural network based on the general price index showed the network has good performance in predicting stock returns and the variables are adjusted based on the general price index, variables, net profit margin, return on assets, current ratio, asset turnover ratio and fixed asset turnover ratio, respectively, are of the greatest importance and impact. Compared to predict stock returns based on financial ratios based on historical cost and ratios adjusted for general price index based on the least squares regression and neural network by using root mean square error performance measures (RMSE), the mean absolute error (MAE) and mean absolute percentage error (MAPE) showed expected stock returns based on financial ratios based on the general price index to predict stock returns financial ratios based on historical cost (in both methods, least squares regression and neural networks), is preferred. The first hypothesis (prediction of stock returns using financial ratios based on the projected cost of the modified historical cost is more accurate) confirmed. The neural network predicts stock returns based on the financial ratios of least squares regression method (based on the historical cost basis and general price index), is preferred. The second hypothesis (prediction of stock returns using financial ratios
ผลและอภิปรายผลIn this study the prediction of stock returns using financial ratios based on historical cost compared to modified expense (accounting for inflation) was studied with neural network approach. Results Descriptive statistics for the variables showed that the average return on equity in the companies surveyed is 29.168. As observed ratios adjusted based on the general price index is lower than the ratios have been modified. For example, the average return on average assets and return on assets adjusted to 0.112 based on the general price index that reflects the 0.094 is that the denominator of this ratio are assets based on the general price index adjustment and have been updated this has led to reduced yields and adjusted. The results predict stock returns based on financial ratios based on historical cost through least squares regression shows between return on equity and return on assets with fixed asset turnover ratio is negative and significant. Between stock returns and the current ratio, asset turnover ratio and the ratio of net income to sales, there is a significant positive relationship. The results predict stock returns based on financial ratios adjusted for general price index by least squares regression showed that the current ratio-adjusted stock returns, asset turnover ratio, the ratio of net income to sales and return on assets and a significant positive relationship exists. The negative correlation between stock returns and fixed asset turnover ratio is significant. The results predict stock returns based on financial ratios based on historical cost through neural network showed the network has good performance in predicting stock returns variable and fixed asset turnover ratio, asset turnover ratio, return on assets, current ratio and net profit margins, respectively, are of the greatest importance and impact. The results predict stock returns based on financial ratios adjusted through neural network based on the general price index showed the network has good performance in predicting stock returns and the variables are adjusted based on the general price index, variables, net profit margin, return on assets, current ratio, asset turnover ratio and fixed asset turnover ratio, respectively, are of the greatest importance and impact. Compared to predict stock returns based on financial ratios based on historical cost and ratios adjusted for general price index based on the least squares regression and neural network by using root mean square error performance measures (RMSE), the mean absolute error (MAE) and mean absolute percentage error (MAPE) showed expected stock returns based on financial ratios based on the general price index to predict stock returns financial ratios based on historical cost (in both methods, least squares regression and neural networks), is preferred. The first hypothesis (prediction of stock returns using financial ratios based on the projected cost of the modified historical cost is more accurate) confirmed. The neural network predicts stock returns based on the financial ratios of least squares regression method (based on the historical cost basis and general price index), is preferred. The second hypothesis (prediction of stock returns using financial ratios
การแปล กรุณารอสักครู่..
