Customer Analysis And Sales Forecast Research For Professional Theaters

Culture industry development has been regarded as a national strategy nowadays and one of its core components is theater.As the culture symbol of a city or a district,theater plays an important role for audience performing arts appreciation and education.Therefore,theater marketing strategies are critical in its management and also become an important artistic marketing research direction.Theater marketing is a kind of experience marketing which focuses on audience behavior and loyalty.Thus,this paper takes S theater as the research object.From the three perspectives of brand sponsor,customer relations and theater box office,an effective method to improve the theater marketing is established and high profit marketing strategy for professional theaters is studied by the deep learning theory.The main research contents and contributions of this paper include as follows:1)Research on the mechanism to enhance the effect of brand sponsorship.An evaluation model for theater customer value is researched deeply and established,dividing the measurement dimensions of theater customer value into three aspects:current value,potential value and loyalty.Through the analysis and mining of historical consumption data of theater customers,a high-relevance and high-distinguishable segmentation value indicators is proposed.According to the distribution characteristics of theater customer value,a theater customer segmentation model based on improved Cluster GAN is constructed.By the adversarial learning of generator and discriminator in the training period to fit the probability distribution of customer value data and using the encoder to reconstruct a probability distribution space with same distribution of customer value data,the cluster segmentation target for theater customers in the latent space is completed.Aiming at the segmented theater customer groups,the different characteristics of different groups in theater consumption behavior is analyzed deeply and an objective basis or theater operators in the process of seeking sponsorship is provided.2)Research on the mechanism to strengthen the relationship of theater customers.The characteristics of theater audience transactions and the factors of customer churn are analyzed,and the definition criteria of theater customer churn is proposed.The objective factors of customer churn and the audience's consumption behavior are discussed.Based on these factors,the audience features such as basic characteristics,historical consumption characteristics,interest characteristics,customer group characteristics and consumption dynamic behavior characteristics are proposed.In terms of model construction,the ability of long and short-term memory networks to process temporal feature information was discussed and analyzed,and a theater audience loss prediction model based on multi-layer long and short-term memory networks was proposed.In terms of model construction,the ability of Long Short Term Memory network in processing temporal characteristic information is discussed and analyzed and a prediction model of theater customer churn based on multi-layer LSTM network is proposed.The model deeply explores the hidden nonlinear relationship between the dynamic time series characteristics and static characteristics of theater customers,which improves the ability of predicting the theater audience churn and provides guarantee for the retention strategies for the customers.3)Research on the mechanism to improve the accuracy of performance sales forecast.By analysis of the S theater's sales status,the five forecast targets of theater sales are defined,namely performance box office,sale time,seat occupancy rate,best popular seat and best performance.The forecast targets are turned into the problem of predicting the following three tasks,namely the next seat sold time interval,seat category and customer level.According to the ticket sales rules and the show content characteristics of S theater,the influencing factors to theater marketing objectively are analyzed based on the aspects of performance content static characteristics,performance price static characteristics,performance time static characteristics and performance sales dynamic characteristics.A theater sales forecast model based on a multitask mechanism is constructed,by using the Gated Recurrent Unit to obtain hidden related features in theater sales sequence characteristics and introducing attention mechanism to extract the key feature information,improving the ability of predict performance sales.