ARIMA Model Analyzes the Tendency and Challenges of Intelligent Marketing in the Era of Digitalization

Abstract The popularization of various types of information technology and equipment will inevitably affect the current marketing situation, the proportion of data technology in product marketing is increasingly aggravated. This paper points out the benefits of big data technology added to the marketing link, and outlines the main implementation methods of the current smart marketing. Establish the sales marketing forecasting model based on ARIMA, use the data time series chart as well as the DF test to carry out a smooth test on the time series of product sales data, draw the seasonal difference after the time series chart, and obtain the difference order. The AIC criterion was applied to determine the model parameters, and the sales forecast results based on the ARIMA model were compared with the actual sales volume. In order to optimize the sales forecasting results of ARIMA model, a combined forecasting model of ARIMA model and BP model is established. Analyze the construction premise of the combined forecasting model and carry out the combined model prediction. Comparing the relative error percentages of the prediction results of ARIMA-BP model, ARIMA model, and BP model, which are 2.948%, 9.045%, and 7.233%, respectively, the combination model reflects a better effect, and the constructed ARIMA-BP combination prediction model is able to optimize the prediction accuracy of the ARIMA model, which is more convenient to carry out the analysis of the product wisdom marketing.