Modeling Analysis of User Consumption Data of Internet Culture Industry in the Context of Emerging Technologies

In the context of the prosperity of the Internet economy and the progress of the cultural industry, the Internet cultural industry is expanding in a big way based on the economic development of the cultural industry, supported by the Internet technology platform and serving the cultural and spiritual experience of consumers. The Internet cultural industry is gradually becoming the backbone of the country’s comprehensive strength for the amount. The Internet to boost the development of China’s cultural industry and cultural export to the outside has become a general trend. In 2018 China’s Internet cultural industry added value accounted for the first time exceeded 3 trillion of GDP, and the added value of the Internet cultural industry in 2021 further increased to 6.63% of GDP. Thus, it can be seen that the share of Wen Internet culture industry in China’s GDP is a steady upward trend and the future development is clear. This paper uses the vector autoregressive model to model and analyze the cultural consumption data of Internet users, and the analysis results show that in 2021, China’s Internet cultural consumption mainly tends to music and video, games, literature, animation, and online cultural content information services in five areas, among which games account for 36.52% of the total Internet cultural consumption. Due to the improvement of China’s comprehensive strength, the difference in CPI of Internet culture consumption between urban and rural groups in China is 11, and the gap is gradually narrowing, and the countryside will be a big new market in the future. The Internet group culture user group as a whole tends to be younger, with the student group accounting for 26% of the user group. Based on the analysis of this study, it is concluded that Internet cultural consumption tends to be young, cultural innovation is the main driving force to promote consumption, and future development is unstoppable.


Introduction
Cultural consumption is both the process of exchange value realization and the beginning of content value regeneration.In traditional linear cultural and economic activities, consumption means the end of economic activity.The Internet has transformed the identity of the "end" of cultural consumption into the "beginning" of reproduction and re-consumption.The Internet cultural economy is a sharing economy that follows the economic law of diminishing marginal costs and increasing marginal benefits.The consumption of cultural content in cyberspace emphasizes the realization of self-worth, spiritual experience, and interactive sharing.It is also because of the active and sharing characteristics of online cultural consumption that audiences unconsciously participate in cultural content reproduction, re-promotion, and re-consumption activities in the process of communicating and sharing cultural content goods.This circular audience participation in cultural content production, promotion, consumption and re-production, re-promotion, and re-consumption activities constitutes the large ecosystem of the network cultural economy.At the same time, the network cultural economy breaks through the weaknesses of traditional cultural consumption in terms of time and space means of payment, and income limitations subvert the economic laws of the traditional cultural industry and brings a wider audience and more experiences to cultural consumption.
Literature [1][2][3] discusses that intangible cultural promotion is the guarantee for cultural content goods to achieve value and exchange value and that compared with traditional cultural marketing, Internet cultural content goods promotion emphasizes "soft" marketing information, interactive marketing communication, and integration of all media.Literature [4] also suggests that cultural marketing activities in the Internet environment tend to be more intangible, and physical, cultural products or services often need to obtain the audience's attention through the dissemination of intangible cultural contents, cultural sentiments, cultural concepts, corporate culture, cultural experiences, etc. Internet culture does not directly present the value of cultural content products or services to consumers; the commodity value needs to be discovered and exchanged by consumers on their own initiative.The literature ¡Error!No se encuentra el origen de la referencia.-7] suggests that the interactive nature of Internet culture marketing is the core of its competitiveness compared to traditional "broadcast and push" culture marketing.In a cultural market environment with a wide variety of products, the key to successful cultural marketing is how to promote different cultural content products for different audiences' deconstruction characteristics [8].Internet cultural marketing technology can achieve accurate information collection, analysis, and pushing, and can use different means to promote cultural content goods according to different audiences' different habits, which is the most suitable means to promote cultural content goods that are both subjective and personalized [9][10][11][12][13].The fragmented cultural content market and fragmented audiences together constitute the new vitality of Internet cultural content goods, and word-of-mouth becomes the most important basis for measuring quality [14][15].The literature [16] argues that human demand and consumption of cultural goods are not irrelevant but an inevitable requirement to improve their cultural quality and literacy.With the development of social production and the division of labor, the cultural industry is bound to be separated from the production of cultural goods, and the number of cultural goods owned will become a measure of wealth and poverty of individuals.According to the literature [17][18], the needs of consumers are the regulator of all production goals, and with the advancement of technology, people will pay more and more attention to leisure time, which will have an important impact on the pattern of consumption.Studying cultural consumption from the perspective of the discipline of psychology, literature [19] proposed the theory of consumption hierarchy needs, where he argued that human needs start from the lowest level of physiological needs, followed by the progression of security needs, belonging and love needs, respect needs, knowledge needs, and aesthetic needs, and the highest level of human needs is the need for self-actualization.Based on this, literature ¡Error!No se encuentra el origen de la referencia.proposed the ERG theory, which divides human consumption needs into three categories, namely, needs for survival, needs for interrelationships, and needs for growth and development.Literature [20][21][22][23] showed that: the Internet culture industry, represented by online music, online literature, online games, online animation, online performance, and online video, has become an important force in promoting the rapid development of the cultural industry and the most dynamic area of cultural consumption in China.Literature [24-¡Error!No se encuentra el origen de la referencia.counted that the current scale of Chinese users of online literature, online music, online video, and online games has reached 481 million, an increase of 210 million compared with five years ago; the scale of the online cultural industry reached 3 trillion yuan, an increase of 250%; the number of original literary works on the Internet exceeded the sum of works published in contemporary literature paper media for 20 years; the online video The market size reached 24 billion yuan, with cumulative playback exceeding 100 billion times, which is enough to see the huge consumption market of the Internet culture industry.However, research on the Internet and cultural consumption remains scarce and has only begun to receive attention and focus in the past year or two [25].Academic research on cultural consumption on the Internet is fragmented, lacking, and fragmented.The literature [27][28][29] has studied mobile cultural consumption from a media perspective, mobile cultural consumption from a consumer culture perspective, mobile cultural consumption in a particular place or region, and mobile cultural consumption from a marketing perspective, respectively.However, these studies lack an overall perception of the Internet cultural consumption situation in China under the complex environment of the mobile Internet and lack systematic analysis.
In order to better analyze the change tendency of user's Internet culture consumption, this paper selects three indicators to analyze the change of user's Internet culture consumption user structure, industry structure, and user consumption tendency, based on vector autoregression (VAR) and error correction (ECM), combined with impulse response and variance decomposition to demonstrate the validity of the model, and uses the vector autoregression and error correction model to make predictions for future short-term values.The model is combined with impulse response and variance decomposition to demonstrate the validity of the model.Then the user structure analysis of the consumption index is studied, and the future development direction of Internet culture is explored from its own fluctuation, the difference between urban and rural areas, and the development change of the Internet culture industry to provide decision-making guidance for enterprises.

Internet Culture Consumption
In the concept of Internet cultural consumption, information in the network environment is the most basic cultural content; audiences in Internet culture are first generated specific cultural needs, and then engaged in the network research and network consumption process, but also in the network browsing process, from the information stimulation and research and consumption of cultural needs; again, in the audience to the massive amount of network cultural information access process, retrieval, browsing, categorization and filtering are the main research means; most online cultural consumption behaviors are based on online information research, which is not only price comparison, but also comprehensive access to information related to cultural content goods or services; online cultural consumption behavior is a dynamic interactive process, from online information research to a long time after cultural consumption is completed, such interactive online publishing, sharing and exchange behaviors have always existed and It constitutes a process of information "reproduction" of the audience's consumption of cultural content.By comparing and analyzing the differences between traditional cultural consumption and online cultural consumption, the cultural consumption behavior in the Internet environment is explored in depth.This is shown in Figure 1.Compared to Internet cultural consumption, the direction of traditional cultural consumption is oneway and less time-sensitive.This is mainly manifested in that; traditional cultural consumption is the end point of cultural industry operations, cultural content goods follow a one-way linear direction of production, promotion, and consumption, and it is difficult for consumers to obtain more relevant information before the cultural content goods are produced.This is mainly due to the one-way nature of information dissemination and occlusion, which makes consumers less active in acquiring information.During the traditional cultural consumption period, the sources of information related to cultural content goods are the producers and promoters, and the channels of information acquisition are the traditional media such as TV, newspapers, magazines, and radio, which are one-way and even delayed in information transmission.As shown in Table 1.In the context of the Internet era, the production of cultural content goods determines cultural consumption.Any cultural content goods are presented to the audience only after all the production and post-production processes are completed, and the process of cultural consumption by consumers

Cost structure
Cultural product input, advertising, subscription fees, etc.

Cost structure
Cultural content input, platform maintenance, etc.

Key Partners
Cultural product providers, users, etc. is mostly passive, one-way, and time-delayed, and the information feedback after cultural confusion is also one-way and time-delayed.Cultural consumption in the network environment has changed this situation completely, mainly because of the openness, interactivity, and immediacy of network communication, which brings the convenience of instant information release and access to cultural consumers anytime and anywhere.The immediacy completely changes the delayed nature of cultural consumption, and the interactivity changes the one-way nature of traditional cultural consumption, making online cultural consumption an interactive and multi-directional consumption state.As shown in Figure 2.

Vector autoregressive model
The VAR model was mainly applied to macroeconomics, and when it was first studied, many scholars believed that the accuracy of the model in forecasting was higher than that of the structural equation.Since many complex and tedious equations could not obtain satisfactory results in the 1960s, VAR models were fundamentally different from the previous research methods and provided new ideas and directions to researchers at that time.Therefore, the study of the VAR model is another expression of the macroeconomic model with microscopic ideas, and at the same time, it is not very demanding on the interrelationship between economic variables.In the vector autoregressive model, which is a system of associative equations, there are endogenous and exogenous variables.From the definition point of view, endogenous variables denote the variables to be determined by the model, and exogenous variables are the variables determined by factors outside the model.However, VAR also has shortcomings, such as model estimation requiring a large number of parameters, so in practice, we usually do not choose more than four variables to build the model.Two things need to be clarified when building the VAR model: first, which variables are interrelated, and include all interrelated variables in the VAR model; second, determine the lag order , so that the model can reflect the relationship between variables the vast majority of the relationships between the variables.
Secondly, the biggest problem with this model is that it has no economic basis, so the researchers understood the VAR model as a simplified form and wanted to try to identify the structural form of the model.The VAR model has the following characteristics: 1) The estimation of parameters in the VAR model is relatively easy.And the accuracy of shortterm forecasting is high, but the accuracy of long-term forecasting is not so satisfactory.
2) The form of VAR models is generally generic, and VAR models are not based on economic theory, so researchers can add as many explanatory variables as they want.For example, researchers have recently found that only unidirectional causal variables can be explained as endogenous variables, and this is done to enhance the explanatory power of the model.
3) In general, a reasonable VAR model satisfies the 2 conditions of small size (not more than 4 variables) and reasonable parameter settings.It is better than complex and cumbersome structural equations, especially for short-term forecasting, because the VAR model can reduce the influence of the restrictions implemented in the structural equations.
In general, a VAR (P) model is defined in Equation ( 1): ( Where is the explained variable and is the exogenous explanatory variable, to obtain the matrix, Equation ( 2): (2) is the coefficient matrix of and is the coefficient matrix of .In general, we assume that the error term is independently and identically distributed and is a white noise sequence.
A proper lag order p can accurately reflect the dynamic effect of the model.If the lag order is too large, it will increase the number of parameters to be estimated and then lead to a decrease in the model degrees of freedom; if the lag order is too small, it will not be able to fully reflect the dynamic process of the model, which shows how to determine the lag order of the model is crucial for VAR, and it is necessary to find a balanced relationship between the degrees of freedom and the lags.In general, the statistics of both the Schwartz criterion (SC) and the Akira pool information (AIC) criterion are used to judge, and when the statistics of both are smallest at the same time, the lag order at this time is chosen, Equation (3-4): (3) (4) in the above two equations denotes the number of parameters to be estimated in the model, denotes the number of endogenous variables, is the length of the sample, is the number of exogenous variables, and is the lag order, which needs to obey the log-likelihood value of the multivariate normal (Gaussian) distribution at the same time, Equation ( 5): (5)

Impulse Response Theory
In general situations, researchers do not like to focus on how many coefficients are in each variable in the VAR model, and they are more concerned about the whole system because there are so many coefficients in the VAR model that it would be putting the cart before the horse to dwell on just a few coefficients that are not satisfied; in addition to that assuming that all the connected relationships within the VAR are considered, it is soon found that each coefficient expresses only a part of the relationship and does not reflect the entire dynamic response process, but the researchers wanted to know what effect a change in one variable would have on the other variables in the response path.With this idea advocated, the coefficients in the VAR play a very small role.So the impulse response function was born.The impulse response tracks the impact of a disturbance's shock on other endogenous variables in the current period and in the future, as the shock of a disturbance can always be transmitted to other variables through the lag structure of the VAR model, while the disturbance has contemporaneous correlation, the shock of a disturbance will also affect other variables in the VAR model.
For the other variables in the impact VAR model.As shown in equation ( 6): (6) The assumption , the initial disturbance value , considers the transmission mechanism of the VAR, so the assumption , which is then the specific form of VAR(2), equation ( 7): (7) The recurrence yields , equation (8-10): The derivation of the above equation shows that the disturbance at a certain moment will have an impact on the system's operation.

Variance decomposition theory
Each disturbance brings some unpredictable part to the prediction of .The variance decomposition discusses the ratio of the effect of the orthogonalized residuals on the predicted mean squared deviation to analyze the contribution of each structural shock to the change in the endogenous variables (generally measured by the variance) and then to evaluate the change in the variables when different shock structures go to the shock.It is necessary to give the variance decomposition of each shock to the variables in the VAR model.And the relatively important information about each stochastic perturbation.Comparing the change of this relatively important information over time, the time lag and relative time lag of this variable effect can be estimated, and in general, the decomposition is chosen to end at the fifteenth period because, by this time, the variance decomposition has all stabilized.

Error correction model
The first step in establishing the error correction model is to conduct the co-integration test, i.e., the long-term equilibrium relationship, and only by adding the error correction term on the basis of the co-integration test can the short-term model be established as a way to reflect the short-term fluctuation effect of the model, where the error correction term is in the co-integration test can be obtained by using least squares regression to obtain the mathematical expression, Equation ( 11): The difference is then deformed to obtain Eq. ( 12): Note , indicating the error correction term.
The original hypothesis that the variable is not the cause of the variable is tested, while the regression equation without constraints is required to be estimated using the following two regression models, equation ( 13): (13) The conditional constraint equation, Equation ( 14): ( ) The residual sum of squares RSS of equation ( 13) and the residual sum of squares RSS of equation ( 14) are used to construct the F-statistic, equation ( 15): Where n is the number of sample observations, k is the number of explanatory variables in the unconstrained equation; and m is the number of parameter restrictions.
3 Data statistics and empirical analysis

Data sources
The research data in this paper were obtained from the Report on the Development of China's Internet Culture Industry published by the China Bureau of Statistics [30].

Occupational structure of Internet culture consumption users
From the statistics of the China Bureau of Statistics for the past five years, for the composition of Chinese Internet culture consumption groups, 25.1% of the student group and 20.1% of the general working group account for the main composition, and these two occupations contribute 51.7% of the total Internet culture consumption, accounting for half of the total consumption.The details are shown in Figure 3.For the analysis of the occupational structure of Chinese Internet culture consumption users, the VAR model is established when the user occupational structure needs to be clearly two variables placed in the VAR model to determine the lag order , so that the model can reflect the vast majority of the relationship effects between variables Morgan coefficient, the specific calculation results are shown in Table 2.The collation results from Table 2 show that the inverse of the mode of each characteristic root is less than 1, so the VAR system is stable.In order to reflect more clearly the inverse of the mode of the AR characteristic root and the position of the unit circle in one step, the structure of the AR system is derived.As shown in Figure 4. From Figure 4, it is found that the inverse of the mode of all eigenvalue roots is within the unit circle, so the Internet culture consumption user structure in VAR (2) model is stable.

Comparison of Urban and Rural Residents' Internet Culture Consumption Index
Since there are differences in living standards, consumption structures, and life rhythms between people in urban and rural areas, in order to explore the changes in the level of Internet culture consumption between urban and rural residents, we analyze the data on the consumption spending of residents from 2015 to 2021 nationwide.As shown in Table 3.In summary, the above analysis shows that the consumption of urban and rural users is higher than the rural CPI, and it is found that the gap between them is gradually decreasing and tilting towards the mean line.As shown in Figure 5.

Data Analysis
According to the results of the VAR model estimation, the parameters of the model estimation are only elaborated from the perspective of macro figures, and no conclusion can be drawn on the effect of the model estimation at present, so the difference between the fitted and real values of the model should be calculated based on the estimated parameters of the model, and if the difference is small, it means that the established occupational structure model is reasonable and stable.
It is found that the response path of users' Internet culture consumption index to itself has been a positive response within the first 5 periods, and it reaches the maximum shock of 0.073 around the 2nd period, generally speaking, the perturbation of itself to itself will have an immediate impact, for example, the increase of the Internet culture consumption index in the previous year may affect the increase in the following year, and the shock gradually starts to decrease after 2 periods, about From the above analysis, we can also conclude that the user consumption index has a positive influence on itself throughout the period, and the time lag is relatively long.And the time lag is relatively long, reaching a maximum in a subsequent period and finally decreasing gradually to 0. The predicted variance of the logged consumer price index is decomposed and collated in a table, and the prediction error of the consumer index is inferred from the data in the table to be mainly from its own impact, which is 100% in the first period and then decreases gradually, decreasing by roughly 1% in each period, indicating that with the passage of time, the user The prediction variance of 1% of the shock of the rate of change of consumption index can be explained by its own change, and the proportion of explanation becomes larger as the number of periods increases, and the rate of change of user consumption index can predict the trend of change in the short term more appropriately.It can be found that the other three variables have a smaller impact on the civil consumption index, and in order to analyze this result more intuitively, the variables are error corrected, as shown in Table 4.The plot of the variance decomposition of each variable against the CPI.The blue curve is the impulse response path curve, and the red curve corresponds to the confidence interval of the function plus or minus two times the standard deviation and sets the response function to 15 periods.First, the graph on the left side of the first row is analyzed, indicating the response of the consumer price index to itself with the path of change.It is found that the response path of the consumer price index to itself has been a positive response within the first 15 periods, and it reaches the maximum shock of 0.073 in about the second period.Generally speaking, the perturbation of itself will have an immediate impact.For example, the increase of the consumer price index in the previous year may affect the increase in the following year, and the shock gradually starts to decrease after 2 periods, and it probably weakens to 0 in about 17 periods.The analysis also leads to the conclusion that the CPI has a positive impact on itself during the whole period, with a long time lag, reaching its maximum in the following period and finally gradually weakening to 0.0073.As shown in Figure 7, the CPI has a positive effect on itself throughout the period, with a relatively long time lag, reaching a maximum in the subsequent period and decreasing to 0.

Conclusion
The Internet and the cultural industry are a mutually complementary relationship.The cultural industry in the Internet environment gradually realizes great integration, which involves not only the integration of upstream and downstream industry chains but also the horizontal integration of different industries such as literature, games, film and television, and animation, which are interconnected with each other and constitute an important link in the new industrial chain of the cultural industry.From the modeling analysis of user consumption data, this paper concludes that the Internet is an innovative ecosystem with a strong agglomeration effect, which gives rise to a new industrial form with a high degree of inclusiveness.The Internet and the cultural industry's "online and offline" and "free and paid" consumption models have broken the boundaries of space, making the traditional cultural consumption model subversive and creating a larger and broader market with the support of new Internet technologies.Market.This model helps to build a pan-Internet and pancultural industry ecosystem, which is important for the practice of China's cultural power strategy.

Figure 1 .
Figure 1.Internet and traditional cultural consumption patterns

Figure 2 .
Figure 2. Relationship between consumption motivation and consumption intention

Figure 3 .
Figure 3. Internet culture consumption structure

Figure 5 .
Figure 5. Urban and rural users' online cultural consumption

Figure 6 .
Figure 6.Total consumption of online cultural products

Table 1 .
Comparative analysis of traditional and online cultural consumption

Table 2 .
System stability structure of VAR

Table 3 .
Comparison of National Urban and Rural Internet Culture Consumption Index

Table 4 .
The variance decomposition values of Lncpi