Analysing the nexus between artificial neural networks and ARIMA models in predicting customer lifetime value (CLV) for complex development of society and industrial activities

Today, the importance of customer relationship is not hidden from anyone, and predicting the value of customer life can help organizations to create an optimal relationship with their customers. The concept of industrial society represents a symbiosis between social and industrial activities using mass-production technologies. A sustainable CRM approach can generate significant benefits for the development of the textile industry. This paper compares ARIMA and neural network models in predicting customer lifetime value. The time-domain of the research is related to the year 2021 in the Lojoor company. To identify the variables needed to predict the value of customer longevity, experts in this field and university professors were used through descriptive survey method and using databases to collect other data. After collecting the data, the required variables were first identified by the Delphi method and then the databases were analysed using the artificial neural network method and the ARIMA model, for which MATLAB software was used. The results showed that both ARIMA and artificial neural network models can be used to predict customer lifetime value. In the case of the artificial neural network, it was observed that in addition to better prediction of the relationship between variables, which assumes them to be nonlinear, the artificial neural network model also performed better in terms of prediction results. In total, the values of MAPE error are 10.3% and MSE error is 11.6% for the neural network model. The neural network model is acceptable.