Screening method of public satisfaction influence factors

Abstract

The invention relates to the technical field of data processing, in particular to a public satisfaction influence factor screening method, which comprises the following steps: acquiring public satisfaction monitoring data; preprocessing the public satisfaction monitoring data to obtain target analysis data; determining an independent variable index and a control variable index; and based on the independent variable index and the control variable index, adopting an HLM multilayer linear regression model to screen the target analysis data so as to obtain influence factors of the public satisfaction. Aiming at the problems that as the public satisfaction is used as hierarchical nested data, influence factors of the public satisfaction cover multiple levels of individuals, places and countries, the crowd satisfaction in the same place is not independent, and a classical regression method cannot be used for analysis, an HLM method is adopted for regression analysis, the influence factors are screened, and contribution of each level of factors to the public satisfaction is recognized; false positive results and key influence factors are effectively prevented from being missed, and an accurate analysis scheme is provided for public satisfaction influence factor research.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and in particular to a public satisfaction influence factor screening method. BACKGROUND

[0002] Under the current political environment of building a service-oriented government, public satisfaction is an important measure, and the government and researchers want to know how to improve public satisfaction, which requires screening all possible influence factors of public satisfaction, screening key influence factors, and controlling these key influence factors to effectively improve public satisfaction. The usual way is to perform regression analysis on public satisfaction, taking influence factors as independent variables and public satisfaction as dependent variables, establishing a regression equation, and the variable with a significant regression coefficient is the key influence factor, and the regression coefficient size represents the influence degree of the variable on public satisfaction.

[0003] However, public satisfaction is a typical hierarchical nested data, for example, the influence factors in the field of public medical and health services are more complex, which not only includes individual level influence factors (such as gender, age, education level, etc.), but also includes local level influence factors (such as local investment, output, customs, etc.), and even national level factors (such as policy, mechanism, etc.) influence. For public satisfaction as hierarchical nested data, its influence factors cover individual, local, and national levels, resulting in that the same local population satisfaction is not independent, and the classical regression method cannot be used for analysis.

[0004] Therefore, the present application provides a public satisfaction influence factor screening method to solve the above problems. SUMMARY

[0005] In view of the above problems, in order to overcome the defects of the prior art, the present application provides a public satisfaction influence factor screening method to solve the above problems that public satisfaction is hierarchical nested data, and its influence factors cover individual, local, and national levels, resulting in that the same local population satisfaction is not independent, and the classical regression method cannot be used for analysis.

[0006] In order to achieve the above purpose, the technical scheme adopted by the present application is as follows:

[0007] A public satisfaction influence factor screening method, comprising:

[0008] obtaining public satisfaction monitoring data;

[0009] preprocessing the public satisfaction monitoring data to obtain target analysis data;

[0010] determining independent variable indicators and control variable indicators;

[0011] Based on the independent variable index and the control variable index, the target analysis data is filtered by using a HLM multilevel linear regression model to obtain the influencing factors of the public satisfaction.

[0012] Preferably, the filtering of the target analysis data based on the independent variable index and the control variable index by using the HLM multilevel linear regression model to obtain the influencing factors of the public satisfaction comprises: determining a nested structure; preprocessing the target analysis data to obtain preprocessed data; establishing an empty model; constructing a filtering model based on the empty model; and processing the target analysis data based on the filtering model to obtain the influencing factors of the public satisfaction.

[0013] Preferably, the preprocessing of the target analysis data to obtain the preprocessed data comprises: centering the target analysis data corresponding to the individual layer variable to obtain the preprocessed data.

[0014] Preferably, the establishment of the empty model comprises: formula initialization of the first layer model, the second layer model and the merged model; and determination of the intra-group correlation coefficient based on the preprocessed data; and when the intra-group correlation coefficient is greater than a preset value, it is determined that the target analysis data is processed by using the filtering model.

[0015] Preferably, the formula initialization of the first layer model, the second layer model and the merged model comprises: the initialization formula of the first layer model is: ,

[0016] wherein, is the satisfaction of the first layer, is the total independent variable index, is the intra-group variance;

[0017] the initialization formula of the second layer model is: ,

[0018] wherein, is the inter-group variance, is the independent variable index of the jth independent variable index, is the independent variable index;

[0019] the initialization formula of the merged model is: ,

[0020] wherein, is the satisfaction, is the independent variable index, is the inter-group variance, is the intra-group variance, i represents the ith control variable index, and j represents the jth independent variable index.

[0021] Preferably, the determining the intragroup correlation coefficient according to the pretreatment data comprises:

[0022] Intragroup correlation coefficient The calculation formula is:

[0023] Wherein, is the intergroup variance, is the intragroup variance.

[0024] Preferably, the preset value is 5.9%.

[0025] Preferably, the constructing the screening model based on the empty model comprises: constructing a random intercept model and a random slope model; and optimizing the random intercept model and the random slope model according to the constraint conditions of key parameters in the random intercept model and the random slope model to obtain an optimized screening model.

[0026] Preferably, the formula of the random intercept model is:

[0027] Wherein, is the satisfaction degree, is the independent variable influence index, is the fixed effect index, is the data value corresponding to the subdivision index of the i th control variable index and the j th independent variable index in the first layer; is the intergroup variance, which is also the random effect index; is the intragroup variance.

[0028] Preferably, the formula of the random slope model is:

[0029]

[0030] Wherein, is the satisfaction degree, is the independent variable influence index, is the fixed effect index, is the data value corresponding to the subdivision index of the i th control variable index and the j th independent variable index in the first layer, is the per capita health and health expenditure corresponding influence index, is the factor value corresponding to the j th independent variable, is the weight parameter; is the intergroup variance, which is also the random effect index; is the random slope variance, is the intragroup variance.

[0031] The beneficial effects of the present application are:

[0032] ​​​In order to screen out the key influencing factors, the traditional method is to perform regression analysis on the public satisfaction, taking the influencing factors as the independent variables and the public satisfaction as the dependent variable, to establish a regression equation, and the variable with a significant regression coefficient is the key influencing factor, and the size of the regression coefficient represents the influence degree on the public satisfaction. However, the public satisfaction is hierarchical nested data, and the influencing factors cover multiple levels of individuals, places and countries, so that the satisfaction of the same place population is not independent, and the classical regression method cannot be used for analysis.

[0033] The process of screening the influencing factors of the public satisfaction by adopting the HLM multilevel linear regression model is as follows: based on the independent variable indexes and the control variable indexes, the HLM multilevel linear regression model is adopted to screen and process the target analysis data, so as to obtain the influencing factors of the public satisfaction. When the HLM multilevel linear regression model is adopted, the nested structure is determined; the target analysis data is preprocessed to obtain the pretreated data; the null model is established; the screening model is constructed based on the null model; and the target analysis data is processed according to the screening model, so as to obtain the influencing factors of the public satisfaction. Through the above manner, the HLM multilevel linear regression model can be adopted to identify the contribution of each layer of influencing factors to the public satisfaction, avoid false positives or omission of key influencing factors, provide an accurate analysis scheme for the influencing factor research of the public satisfaction, and help to solve the problem that the public satisfaction is hierarchical nested data, the influencing factors cover multiple levels of individuals, places and countries, the satisfaction of the same place population is not independent, and the classical regression method cannot be used for analysis. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 The screening method for the influencing factors of the public satisfaction is shown in the schematic flowchart. DETAILED DESCRIPTION

[0035] The specific implementation of the present application will be described below with reference to the accompanying drawings. Figure 1 The embodiments of the present application will be described in detail. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present application, and are not intended to limit the protection scope of the present application.

[0036] The screening method for the influencing factors of the public satisfaction is shown in the schematic flowchart. Figure 1 The screening method for the influencing factors of the public satisfaction is shown in the schematic flowchart.

[0037] Step S11: obtaining public satisfaction monitoring data;

[0038] Step S12: preprocessing the public satisfaction monitoring data to obtain target analysis data;

[0039] Specifically, one preprocessing method is to sample the public satisfaction monitoring data by adopting hierarchical second-order sampling method to obtain first sampling data; and the target analysis data is obtained by screening the first sampling data.

[0040] Another pre-processing mode is: setting a screening condition according to a data requirement to be obtained, for example, a keyword, a time, a region and the like, directly performing screening in one step to obtain target analysis data.

[0041] Step S13: determining independent variable indexes and control variable indexes.

[0042] For example, when screening influencing factors of medical and health services, the control variable indexes include genders, ages, education levels and income levels of the public; and the corresponding independent variables are objective performance of the medical and health services, and the independent variable indexes specifically include three types of indexes, namely, input, output and effect.

[0043] Step S14: based on the independent variable indexes and the control variable indexes, screening and processing the target analysis data by using a HLM multilevel linear regression model to obtain influencing factors of the public satisfaction.

[0044] Based on the above mode, in order to screen out key influencing factors of the public satisfaction, compared with a traditional method of taking the influencing factors as independent variables, taking the public satisfaction as a dependent variable, establishing a regression equation, and taking a variable with a significant regression coefficient as a key influencing factor, and taking a regression coefficient as a degree of influence on the public satisfaction, the present application screens and processes the target analysis data by using a HLM multilevel linear regression model based on the independent variable indexes and the control variable indexes to obtain the influencing factors of the public satisfaction; the present application can identify a contribution of each layer of the influencing factors to the public satisfaction, avoids false positives or omission of the key influencing factors, and is helpful to solve a problem that the public satisfaction is hierarchical and nested data, the influencing factors thereof cover individual, local and national levels, the satisfaction of the same local population is not independent, and a classical regression method cannot be used for analysis.

[0045] Further, in an embodiment of the present application, the screening and processing of the target analysis data by using the HLM multilevel linear regression model based on the independent variable indexes and the control variable indexes to obtain the influencing factors of the public satisfaction includes: determining a nested structure; preprocessing the target analysis data to obtain preprocessed data; establishing an empty model; constructing a screening model based on the empty model; and processing the target analysis data according to the screening model to obtain the influencing factors of the public satisfaction.

[0046] The determination of the nested structure includes: determining a hierarchical relationship of the data, and determining and storing data of each level to a corresponding data storage list. For example, the hierarchical relationship is from the public individual to the performance of the medical and health field at the city level.

[0047] When the screening model is constructed based on the empty model, all significant variables are integrated, and a fitting degree of the screening model is optimized.

[0048] Further, the preprocessing of the target analysis data to obtain preprocessed data comprises: centering the target analysis data corresponding to the individual layer variable to obtain the preprocessed data.

[0049] Specifically, the data corresponding to the first layer independent variable index is subjected to group mean centering or total mean centering to distinguish the within-group effect and the between-group effect.

[0050] In an embodiment of the present application, the establishment of the empty model comprises: formula initialization of the first layer model, the second layer model and the combined model; and determination of the within-group correlation coefficient according to the preprocessed data; and when the within-group correlation coefficient is greater than a preset value, it is determined that the target analysis data is processed by using the screening model.

[0051] Specifically, the formula initialization of the first layer model, the second layer model and the combined model comprises:

[0052] The initialization formula of the first layer model is: ,

[0053] wherein, is the satisfaction degree of the first layer, is the total independent variable impact index, is the within-group variance, i represents different individuals in the first layer, that is, the control variable index, and j represents different groups in the second layer, that is, the subdivision index of each independent variable index.

[0054] The initialization formula of the second layer model is: ,

[0055] wherein, is the between-group variance, is the independent variable impact index of the jth independent variable, is the independent variable impact index.

[0056] The initialization formula of the combined model is: ,

[0057] wherein, is the satisfaction degree, is the independent variable impact index, is the between-group variance, is the within-group variance, i represents the ith control variable index, and j represents the subdivision index of the jth independent variable index.

[0058] Further, the determination of the within-group correlation coefficient according to the preprocessed data comprises:

[0059] The calculation formula of the within-group correlation coefficient is: ,

[0060] wherein, is the inter-group variance, is the intra-group variance.

[0061] Preferably, the preset value is 5.9%.

[0062] Through the above embodiment, the present application can reasonably determine the empty model, so that when subsequent screening of the influencing factors of satisfaction, the key influencing factors of satisfaction are comprehensively and objectively screened out.

[0063] In an embodiment of the present application, the screening model based on the empty model is constructed, comprising: constructing a random intercept model and a random slope model; optimizing the random intercept model and the random slope model according to the constraint conditions of the key parameters in the random intercept model and the random slope model to obtain an optimized screening model.

[0064] The formula of the random intercept model is: ,

[0065] Wherein, is the satisfaction, is the independent variable influence index, is the fixed effect index, is the data value corresponding to the first layer i-th control variable index and j-th independent variable index subdivision index; is the inter-group variance, which is also a random effect index; is the intra-group variance.

[0066] Preferably, is the fixed effect index, indicating the average influence of the control variable of the individual layer on the result; is the inter-group variance, which is also a random effect index, indicating whether the inter-group variation of the intercept is significant.

[0067] The formula of the random slope model is: ,

[0068] Wherein, is the satisfaction, is the independent variable influence index, is the fixed effect index, is the data value corresponding to the first layer i-th control variable index and j-th independent variable index subdivision index, is the influence index corresponding to the per capita health and health expenditure, is the factor value corresponding to the j-th independent variable, is the weight parameter; is the inter-group variance, which is also a random effect index; is the random slope variance, is the intra-group variance.

[0069] Preferably, The cross-level interaction item is formed, and the adjustment effect of the second layer independent variable index on the relationship between the first layer control variable index is verified. The first layer is a public individual, and the second layer is a prefecture-level city.

[0070] Through the setting mode of the embodiment, the application can construct a comprehensive and reasonable screening model on the basis of the empty model combined with the independent variable index and the control variable index in the medical and health service data, so as to obtain the influencing factors of public satisfaction and reasonably evaluate the satisfaction of the public on the medical and health service.

[0071] In an embodiment of the application, the related experimental content of the application includes:

[0072] In the field of public medical and health services, the objective performance factors affecting public satisfaction mainly include three aspects of input, output and effect.

[0073] The dependent variable is the satisfaction of the public on the medical and health service. The data is derived from the 2022 national public service quality monitoring data. The 2022 national public service quality monitoring adopts a stratified two-stage sampling method, and 40161 public data in 120 prefecture-level cities in 31 provinces (autonomous regions and municipalities) are extracted.

[0074] Considering the acquisition of medical service performance index data, 3374 valid samples covering 102 prefecture-level cities in 30 provinces (autonomous regions and municipalities) are extracted from the 2022 national public service quality monitoring data.

[0075] The independent variable index is the objective performance of medical and health services, specifically including three types of indexes of input, output and effect (see Table 1). The data is derived from the statistical yearbook of each province and the official website of the health commission.

[0076] The control variable index is the gender, age, education level and income level of the public, and the data is derived from the 2022 national public service quality monitoring data.

[0077] Table 1 Medical and health service objective performance index variable

[0078]

[0079] In this study, the HLM multilevel linear regression model is used to explore the performance of the medical and health field from two levels of public individuals (the first level) and prefecture-level cities (the second level). Among them, the individual variables in the first level include gender, age, education and income level, and the indexes in the second level include per capita health expenditure, thousand people hospital bed number, thousand people licensed (assistant) physician number, population birth rate and per capita visit frequency. The built model is as follows:

[0080] First layer: medical and health service satisfaction

[0081] ,

[0082] In the second layer:

[0083]

[0084] ;

[0085] The mixed model is obtained by superimposing the first layer and the second layer as follows:

[0086]

[0087] The research results are as follows:

[0088] According to the analysis steps of the double-layer linear regression model, first, an empty model without any independent variables is established to determine the size of the explanatory power of individual-level differences and local-level differences on medical and health service satisfaction.

[0089] The first layer: medical and health service satisfaction = ; The second layer: .

[0090] The results of the empty model are shown in Table 3. The variance component in the first layer is 1.51262, and the variance component in the second layer is 0.37060. The variance in the second layer accounts for 19.7% of the total variance (> 5.9%), which meets the conditions for hierarchical linear regression analysis, and the next step of analysis can be conducted.

[0091] Table 3 Differences within the data of the two layers

[0092]

[0093] According to the analysis results of the above empty model, the corresponding data of the variable indicators of the first layer and the second layer are then included in the model to study the impact of all variables on medical and health service satisfaction in the hierarchical model. The objective performance indicators of input, output, and effect are gradually included in the model, which are model 1 (input), model 2 (output), model 3 (effect), model 4 (input-output), model 5, and model 6 (input-output-effect). The fixed effect estimation results are shown in Table 4.

[0094] The results of Model 1 show that, at the prefecture level, the per capita health expenditure has a positive effect on the satisfaction of medical and health services, and the effect is statistically significant (P<0.05). At the individual level, gender and income level have a significant effect on the satisfaction of medical and health services, while age and education level have no significant effect. For gender, men are more satisfied with medical and health services than women. For income level, the higher the monthly household income, the higher the satisfaction of medical and health services.

[0095] Table 4 Results of the effect of each variable on the satisfaction of medical and health services Fixed effect estimation table

[0096]

[0097]

[0098] Note: The results are variable coefficients; T values are in parentheses; ***, **, * represent significance at the 1%, 5%, and 10% levels, respectively.

[0099] The results of Model 2 show that, at the prefecture level, the number of hospital beds per thousand people has a positive effect on the satisfaction of medical and health services, and the effect is statistically significant (P<0.05). The number of practicing (assistant) physicians per thousand people has no significant effect on the satisfaction of medical and health services. At the individual level, similar to the results of Model 1, gender and income level have a significant effect on the satisfaction of medical and health services, while age and education level have no significant effect.

[0100] The results of Model 3 show that, at the prefecture level, the birth rate and the number of visits per person have a positive effect on the satisfaction of medical and health services, and the effect of the number of visits per person is more significant (P<0.01). At the individual level, similar to Models 1 and 2, gender and income level have a significant effect on the satisfaction of medical and health services, while age and education level have no significant effect. However, the significance of the effect of income level on the satisfaction of medical and health services has decreased.

[0101] According to the results of the analysis of the previous three models, the indicators that have a significant impact on the satisfaction of medical and health services are gradually included in the model. In Model 4, input and output indicators are included, and in Model 5, effect indicators are further included. The results of Model 5 show that among the objective performance indicators of the effect category, the number of visits per capita still has a very significant impact on the satisfaction of medical and health services (P<0.01), but the impact of the birth rate on the satisfaction of medical and health services has changed from significant to insignificant. Model 6 further optimizes the model based on the results of Model 5, removing the population birth rate indicator. The final model results show that at the prefecture-level city level, per capita health expenditure, hospital beds per thousand people, and the number of visits per capita have a significant impact on the satisfaction of medical and health services; at the individual level, gender and income level have a significant impact on the satisfaction of medical and health services.

[0102] Finally, the variance component test is used to test the explanatory power of the three types of medical and health service objective performance on satisfaction. The results show that after gradually adding input, input-output, and input-output-effect indicators, the variance component tends to continuously decrease. It can be seen that the objective performance of medical and health services can provide certain explanation for the differences in satisfaction. See Table 5 for details.

[0103] Table 5. Explanation of three types of medical and health service objective performance indicators

[0104]

[0105] (Five) Comparison with simple linear regression method

[0106] According to the above independent variable and dependent variable data, the simple linear regression method is used to find that in the regression equation at the individual level, only income has a significant impact on the satisfaction of medical and health services.

[0107] Table 6.1 Linear regression variance coefficient at the individual level

[0108]

[0109] In the regression equation at the prefecture-level city level, per capita health expenditure, hospital beds per thousand people, and the number of visits per capita have a significant impact on the satisfaction of medical and health services.

[0110] Table 6.2 Linear regression variance coefficient at the prefecture-level city level

[0111]

[0112] Compared with the stratified linear regression method, the simple linear regression method did not identify the impact of gender variable on health service satisfaction. From the explanatory power of the regression equation, the R2 value of the stratified linear regression equation is greater than that of the simple linear regression equation, indicating that the stratified linear regression method has greater explanatory power of the variance of the dependent variable (health service satisfaction) of the selected independent variables.

[0113] Table 7 Explanatory power of simple linear regression equation and stratified linear regression equation

[0114]

[0115] Note: The R2 of stratified linear regression method is calculated according to the research of Kreft, de Leeuw (1998), Singer (1998) and Snijders, Bosker (1999).

[0116] The various implementations of the systems and techniques described above can be realized in digital electronic circuitry, integrated circuitry, specially designed application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0117] It should be noted that in the description of the present application, the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer" and the like terms indicating the direction or positional relationship are based on the direction or positional relationship shown in the drawings, which is only for the convenience of description, and does not indicate or imply that the device or element must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first", "second", "third" are only for the purpose of description, and cannot be understood as indicating or implying relative importance.

[0118] Moreover, it needs to be explained that in the description of the present application, unless explicitly specified and limited, the terms "mounting", "connection", "connecting" should be understood in a broad sense, for example, can be fixed connection, can also be detachable connection, or integrally connected; can be mechanical connection, can also be electrical connection; can be directly connected, can also be indirectly connected through an intermediate medium, can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0119] Program code for carrying out the methods of the present disclosure can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, produces the functions / operations specified in the flowcharts and / or the block diagrams. The program code can be executed entirely on a machine, partially on a machine, partially on a machine as a stand-alone software package, and partially on a remote machine or entirely on a remote machine or server.

[0120] In the context of the present disclosure, a machine-readable medium can be a tangible medium that contains or stores program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disc read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0121] To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0122] The systems and techniques described herein can be implemented in a computing system that includes a back end component, e.g., as a data server, or that includes a middleware component, e.g., an application server, or that includes a front end component, e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described herein, or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication, e.g., a communication network. Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0123] The computer system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0124] So far, the technical solutions of the present application have been described in combination with the preferred embodiments shown in the drawings, but those skilled in the art can easily understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without departing from the principles of the present application, and the technical solutions after these changes or replacements will all fall within the protection scope of the present application.

Claims

1. A method for screening factors influencing public satisfaction, characterized in that, include: Obtain public satisfaction monitoring data; Preprocessing of public satisfaction monitoring data yields target analysis data; Determine the independent variable indicators and control variable indicators; Based on independent and control variables, the HLM multilevel linear regression model was used to screen and process the target analysis data to obtain the influencing factors of public satisfaction.

2. The screening method according to claim 1, characterized in that, The method described above, which uses an HLM (High-Level Linear Regression) model to screen target analysis data based on independent and control variables to obtain factors influencing public satisfaction, includes: determining the nested structure; preprocessing the target analysis data to obtain preprocessed data; establishing an empty model; constructing a screening model based on the empty model; and processing the target analysis data according to the screening model to obtain factors influencing public satisfaction.

3. The screening method according to claim 2, characterized in that, The aforementioned preprocessing of the target analysis data to obtain preprocessed data includes: centering the target analysis data corresponding to individual-level variables to obtain preprocessed data.

4. The screening method according to claim 2, characterized in that, The establishment of the empty model includes: initializing the formulas for the first-level model, the second-level model, and the merged model; determining the intra-group correlation coefficient based on the preprocessed data; and determining that the target analysis data is processed using a screening model when the intra-group correlation coefficient is greater than a preset value.

5. The screening method according to claim 4, characterized in that, The formula initialization of the first-layer model, the second-layer model, and the merged model includes: The initialization formula for the first layer model is: , in, For the first level of satisfaction, The total influence index of the independent variable. Within-group variance; The initialization formula for the second-layer model is: , in, The variance between groups, Let j be the influence index of the independent variable. The influence index is determined by the independent variable. The initialization formula for the merged model is: , in, For satisfaction, The index is influenced by the independent variable. The variance between groups, Here, denoted as within-group variance, i represents the i-th control variable indicator, and j represents the sub-indices of the j-th independent variable indicator.

6. The screening method according to claim 4, characterized in that, The determination of the intragroup correlation coefficient based on preprocessed data includes: Within-group correlation coefficient The calculation formula is: , in, The variance between groups, The variance is the within-group variance.

7. The screening method according to claim 4, characterized in that, The preset value is 5.9%.

8. The screening method according to claim 1, characterized in that, The aforementioned screening model based on an empty model includes: constructing a random intercept model and a random slope model; optimizing the random intercept model and the random slope model according to the constraints of key parameters in the random intercept model and the random slope model to obtain the optimized screening model.

9. The screening method according to claim 8, characterized in that, The formula for the stochastic intercept model is as follows: , in, For satisfaction, The index is influenced by the independent variable. For fixed effects index, These are the data values ​​corresponding to the sub-indicators of the i-th control variable indicator and the j-th independent variable indicator in the first layer; This represents the variance between groups, and is also the random effects index. The variance is the within-group variance.

10. The screening method according to claim 8, characterized in that, The formula for the stochastic slope model is: , in, For satisfaction, The index is influenced by the independent variable. For fixed effects index, These are the data values ​​corresponding to the sub-indicators of the i-th control variable indicator and the j-th independent variable indicator in the first layer. The impact index corresponding to per capita health expenditure. Let j be the factor value corresponding to the j-th independent variable. These are weight parameters; This represents the variance between groups, and is also the random effects index. For the random slope variance, The variance is the within-group variance.

Images