Enterprise Sales Data Processing Method And Market Perception Research In Big Data Environment

With the development of information,sales data is increasing.Because of the importance of the sales data,how to find out the useful sales data is the must for the enterprise.It is urgent for the enterprise to find out the data processing method and make proper use of the sales data.Based on massive amounts of sales data contained in the sales data to predict the future direction of the trend of sales data,this paper provides effective reference information and decision-making for the enterprise marketing decision makers.According to the problems mentioned above,I make some research about data mining method and market forecasting model as follows:(1)Use Hadoop to store data and use MapReduce to extract data.Then import the data into the relational database.According to the demand of data structure in the data mining algorithm and the different data anomalies,I will take different cleaning strategy to clean and standardize the data.Deliver data to the relational database.(2)According to the problems in the data mining algorithm,I come up with the KNN classification algorithm which is based on block overlap k-means clustering.The algorithm is to add a training process to the KNN classification algorithm.It can classify data accurately and quickly.Classify the sales market.Statistic classification results.Compared with the actual data to verify the feasibility and accuracy of the algorithm.(3)Analysis the applicability of various forecasting models for the research content of this paper,and according to the characteristics of this data and the results data characteristics expected,I choosed the suitable forecasting model of ARIMA forecasting model(ARIMA(p,d,q))and grey model,as the basis of the market awareness model.(4)Taking retail sales data for the experimental data,build a model combination ARIMA forecasting model with gray model.Although ARIMA forecasting model can predict short-term sales data,the longer the forecast time prediction accuracy is lower.On the basis of the ARIMA forecasting model,the enterprise use grey model for long-term sales data to predict,allowing businesses to be able to see not just the next six months or a year of sales data and grasp the future trend of the data more accurately.