The purpose of this study is to predict actual sales of sports center by using publicly available weather data, macroeconomic index, stock price index, Internet search volume and sales information of sports center. In addition, the purpose of this study is to find out which analytic algorithms show ...
The purpose of this study is to predict actual sales of sports center by using publicly available weather data, macroeconomic index, stock price index, Internet search volume and sales information of sports center. In addition, the purpose of this study is to find out which analytic algorithms show the most accurate forecasting power, using decision tree, support vector machine, linear regression analysis, and RNN model of deep learning method among the machine learning algorithms used in this study. For the progress of this study, we collected sales data of sports centers located in Busan city. The sales data collected daily data for three years and seven months from January 2015 to July 2018. Based on the collected sales data, data on weather information, stock price index, and internet search volume were collected the same as when the sports center sales occurred. Data were pre-processed for data analysis of this study, and all analyzes were done using open source Python version 3.6. The performance evaluation results on the forecasting power of sales are as follows.
First, to verify the predictive power of the decision tree model, the performance of the decision tree model was evaluated by dividing the whole data set into training data set of 70%, validation set of 10%, and test data set of 20%. The predictive power for the entire data set was 87.5%.
Second, to verify the performance of the support vector machine, the performance of the support vector machine model was evaluated by dividing the entire data set into training data set 70%, validation set 10%, and test data set 20%. As a result of the performance evaluation, the predictive power of the entire data set was 91.3%, but only the predicted result of the sales amount was maintained, and the forecasting of the increase and decrease was not made.
Third, a linear regression model is used to predict sales. For the analysis of the regression model through the sklearn library, the data sets were categorized as 70% validation data set 10% test data set 20%. The forecasting rate for the entire data was 23% accurate.
Fourth, to evaluate the performance of the RNN model, performance evaluation was performed by dividing the train data into 70% and the validation data set 10%, and the test data set 20%. As a result of the performance evaluation, the forecasting rate for the entire data set was 92%.The RNN model, which showed the highest forecasting rate among the final evaluation models, revealed that weather information, stock price index, and internet search volume were the main factors for forecasting sales.
According to the results of this study, managers of sports centers will be able to make efficient decisions in the process of establishing management strategies and marketing strategies through accurate forecasting using data such as weather information, stock price index, and internet search volume.
In conclusion, this study suggests ways to improve the accuracy of sales forecasting of sports center through data analysis and it can be a basic data for efficient strategy decision of sports center operation based on the results of research
The purpose of this study is to predict actual sales of sports center by using publicly available weather data, macroeconomic index, stock price index, Internet search volume and sales information of sports center. In addition, the purpose of this study is to find out which analytic algorithms show the most accurate forecasting power, using decision tree, support vector machine, linear regression analysis, and RNN model of deep learning method among the machine learning algorithms used in this study. For the progress of this study, we collected sales data of sports centers located in Busan city. The sales data collected daily data for three years and seven months from January 2015 to July 2018. Based on the collected sales data, data on weather information, stock price index, and internet search volume were collected the same as when the sports center sales occurred. Data were pre-processed for data analysis of this study, and all analyzes were done using open source Python version 3.6. The performance evaluation results on the forecasting power of sales are as follows.
First, to verify the predictive power of the decision tree model, the performance of the decision tree model was evaluated by dividing the whole data set into training data set of 70%, validation set of 10%, and test data set of 20%. The predictive power for the entire data set was 87.5%.
Second, to verify the performance of the support vector machine, the performance of the support vector machine model was evaluated by dividing the entire data set into training data set 70%, validation set 10%, and test data set 20%. As a result of the performance evaluation, the predictive power of the entire data set was 91.3%, but only the predicted result of the sales amount was maintained, and the forecasting of the increase and decrease was not made.
Third, a linear regression model is used to predict sales. For the analysis of the regression model through the sklearn library, the data sets were categorized as 70% validation data set 10% test data set 20%. The forecasting rate for the entire data was 23% accurate.
Fourth, to evaluate the performance of the RNN model, performance evaluation was performed by dividing the train data into 70% and the validation data set 10%, and the test data set 20%. As a result of the performance evaluation, the forecasting rate for the entire data set was 92%.The RNN model, which showed the highest forecasting rate among the final evaluation models, revealed that weather information, stock price index, and internet search volume were the main factors for forecasting sales.
According to the results of this study, managers of sports centers will be able to make efficient decisions in the process of establishing management strategies and marketing strategies through accurate forecasting using data such as weather information, stock price index, and internet search volume.
In conclusion, this study suggests ways to improve the accuracy of sales forecasting of sports center through data analysis and it can be a basic data for efficient strategy decision of sports center operation based on the results of research
주제어
#머신러닝 딥러닝 스포츠센터 공공데이터 인터넷 검색 데이터 Machine Learning Deep Learning Sports Center Public Data Internet Search Data
※ AI-Helper는 부적절한 답변을 할 수 있습니다.