| The box office returns of the Chinese is increasing year by year.As more and more capitals enter,lots of new theaters are emerging in cities at all levels.This makes the sites selection of new theaters are very important,but the traditional site selection method has little research on quantitative prediction and relies too much on the experience of the evaluators,and it can no longer meet the needs in the competitive market at this stage.The main challenges of the site selection are as follow: The overall box office of the core business districts of large cities is high.However,due to fierce competition,the average box office of theaters is not necessarily high;As for small cities or secondary business districts,although there are few competitors,the lack of supporting facilities and underdeveloped ecology will also affect the box office.So it is necessary to use scientific methods to predict the box office of the new theater.This article hopes to establish a spatial econometric model based on various business indicators and historical box office data of 191 theaters built before 2018 in Beijing to make a reasonable prediction of the box office of new theaters built after 2018.The main research contents are as follows:(1)Descriptive statistics and correlation analysis of the data show that there is a significant positive correlation between the new theater‘s box office and the number of screens,number of seats,average ticket price and other indicators.And through the three-dimensional scatter diagram,we find that the indicators such as the occupancy of the theater show a positive pyramid shape.(2)Using Moran’s I coefficient to do spatial correlation test,we use spatial weight matrix such as K-nearest neighbors spatial weight matrix,spatial weight matrix based on distance threshold and three spatial weight matrixes based on economic data.(3)Using some normal spatial weight matrix and I proposed three matrix based on experience,we establish spatial autoregressive model with maximum likelihood estimation and two-stage least squares estimation to do parameter estimation;At the same time we establish spatial autocorrelation model with maximum likelihood estimation and generalized least squares estimation to do parameter estimation,also with regression residual and prediction accuracy to choose the best spatial autoregressive model and spatial autocorrelation model.(4)Using the 25 Beijing cinemas’ data which built in 2018 as a validation,the final models is spatial autoregressive model using the spatial weight matrix based on the box office per screen and seat returns proposed by me.The prediction accuracy of this model is greatly improved compared with the multiple linear regression model.It is also better than other spatial models under others spatial weight matrix. |
