(三) 資料預處理本研究在資料的前處理將現有資料中每日的成交量最高到最

(三) 資料預處理
本研究在資料的前處理將現有資料中每日的成交量最高到最低，利用十平均分配的運算，將其分類10~1的類別，第10類為最高成交量的分類，反之則第1類則是為最低成交量的分類。
(四) 預測正確率
本研究是將原始資料利用平均分配將成交量分出十等份，在經過訓練模型之後預測出來的結果，將原先分類結果與預測分類進行比對，其正確率公式為:
Accuracy = ((X - P) / 10 ) * 100)/Count)*100
X = 原本分類結果
P = 預測出分類結果
Count = 分類資料個數
最後再將每筆正確率與預測總數經過平均計算出來最後的正確率。

ผลลัพธ์ (อังกฤษ) 2:

(Iii) information on the pretreatment
of the present study data before processing the existing information in the highest to the lowest daily volume, the use of ten evenly distributed computing, to classify the category of 10 to 1, 10 being the best volume classification, and vice versa Category 1 is the lowest volume category.
(Iv) the prediction accuracy
of the present study is the raw data using the average daily volume of distribution will be separated ten equal parts, after training model predicted results, the original classification classification results were compared with the forecast, the correct rate formula is :
Accuracy = ((X - P) / 10) * 100) / Count) * 100
X = original classification
P = predicted classification result
Count = number of classified information
and then finally each correct prediction rate and total average calculation after out of the final accuracy.

ผลลัพธ์ (อังกฤษ) 3:

(three) data preprocessing
This study will present the highest volume to the lowest in the data processing, and use the arithmetic of the average allocation of ten to classify the classification of 10~1, the tenth is the highest turnover, while the first is the lowest.
(4) prediction accuracy
This study is original data the average distribution of the volume is divided into ten equal parts. After training the model predicted results in, the original classification results with the predicted classification comparison, the correct rate formula: =
Accuracy ((x - P) / 10) * 100) /Count *100
X = original classification results
P = forecast the result of the classification
Count = the number of categories of data.