CN107180160A - Public bicycles consumer loyalty degree based on SEM models determines method - Google Patents
Public bicycles consumer loyalty degree based on SEM models determines method Download PDFInfo
- Publication number
- CN107180160A CN107180160A CN201710474219.9A CN201710474219A CN107180160A CN 107180160 A CN107180160 A CN 107180160A CN 201710474219 A CN201710474219 A CN 201710474219A CN 107180160 A CN107180160 A CN 107180160A
- Authority
- CN
- China
- Prior art keywords
- variable
- model
- public bicycles
- mrow
- questionnaire
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000012360 testing method Methods 0.000 claims abstract description 38
- 238000007086 side reaction Methods 0.000 claims abstract description 7
- 230000000694 effects Effects 0.000 claims description 27
- 238000011156 evaluation Methods 0.000 claims description 21
- 238000011835 investigation Methods 0.000 claims description 10
- 238000003556 assay Methods 0.000 claims description 9
- 238000005259 measurement Methods 0.000 claims description 8
- 230000014509 gene expression Effects 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000000903 blocking effect Effects 0.000 abstract description 2
- 238000004458 analytical method Methods 0.000 description 22
- 238000011160 research Methods 0.000 description 17
- 238000007689 inspection Methods 0.000 description 11
- 230000008447 perception Effects 0.000 description 9
- 238000009826 distribution Methods 0.000 description 6
- 238000013525 hypothesis research Methods 0.000 description 6
- 230000033772 system development Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000013480 data collection Methods 0.000 description 3
- 238000013101 initial test Methods 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 239000000203 mixture Substances 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000001364 causal effect Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000001351 cycling effect Effects 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000009472 formulation Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009440 infrastructure construction Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000001568 sexual effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 230000032258 transport Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G06Q50/40—
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16Z—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS, NOT OTHERWISE PROVIDED FOR
- G16Z99/00—Subject matter not provided for in other main groups of this subclass
Abstract
Method is determined the invention discloses a kind of public bicycles consumer loyalty degree based on SEM models, is comprised the following steps：Build the public bicycles consumer loyalty degree initial theory model based on structural equation model；By pre- questionnaire gathered data, KMO values, side reaction coefficient, factor loading, accumulative explained variance are calculated using questionnaire data, questionnaire quality is tested；The polynary kurtosis value and polynary degree of bias value of questionnaire data are calculated, normal distribution-test is carried out to the valid data of acquisition, public bicycles consumer loyalty degree model is tested and corrected；According to model output result, each latent variable is calculated to the value of utility of loyalty and the Satisfaction Index of public bicycles loyalty variation.Quantitative assessment can be carried out to public bicycles consumer loyalty degree and its influence factor using the inventive method, and then public bicycles share rate or even slow-moving traffic share rate can be lifted, so as to make significant contribution in terms of urban traffic blocking is alleviated.
Description
Technical field
Determine method the present invention relates to a kind of public bicycles consumer loyalty degree based on SEM models, for it is public voluntarily
The structure of the structural equation model of car consumer loyalty degree, so as to analyze the method for public bicycles loyalty influence factor, is used for
The research field of user's travel behaviour in transit trip.
Background technology
Public bicycles possess greening and environmental protection, the low advantage of convenient, use cost, are at home and abroad promoted rapidly, into
Slow-moving traffic is promoted for many cities, solve the new measure of last one kilometer problem, but its too fast promotion rate causes big portion
Divide the utilization rate of the public bicycles system in city all relatively low, serious waste of resources.Many cities, which are used, improves infrastructure construction
If mode solve that public bicycles utilization rate is low, the low problem of citizen's participation rate, but the user's request management of lack of targeted
Measure, improvement is not still obvious；But also there is correlative study to find by accurately holding public bicycles user personality, using inclined
The information such as good, formulates public bicycles System Development strategy and Management plan, can effectively lift public bicycles system utilization rate.
Research to public bicycles user's travel behaviour at present, is based primarily upon the progress of the methods such as Discrete Choice Model, right
Individual subscriber characteristic (such as sex, level of education, income, age), scene selection preference etc. carry out proof analysis, study public
The central factor of bicycle user trip preference.Discrete Choice Model is applied to combine that individual subscriber characteristic research is single or part
Influencing factors for demand is difficult the inherent cause and effect many influence factors of user's request and influence to the influence using preference etc.
Mechanism is effectively analyzed, and is lacked where overall aspect dissects attraction of user's continuation using public bicycles, very
Hardly possible helps to form a efficient, with strong points, sustainability system operation and development program.
Public bicycles consumer loyalty degree refers to that user reuses the wish size of public bicycles, lifts user couple
The loyalty of public bicycles has huge impetus to the utilization rate for improving whole system, meanwhile, more can correct guidance system
Construction in a systematic way is set, and reduces operation cost.Either from the perfect angle of public bicycles system Construction, or requirement forecasting with analysis,
The user behavior characteristic research of unification, all can not well " user how to be just ready be used for a long time public bicycles " this
Key problem comprehensively, effectively comb and study, therefore is to accurately hold " how just user to the research of consumer loyalty degree
Be ready be used for a long time " most directly, effective manner.
Structural equation model (SEM models) is widely applied in fields such as social science, business, its research applied
Mainly there are two aspects in direction：One is that can be very good to observe and handle latent variable, that is, is difficult to variable measured directly, such as full
Meaning, sentimental value etc.；Two be that can excavate influence relation that may be present between variable.As can be seen here, for public transportation system,
Structural equation model is equally applicable, but careful, in-depth study not deployed in terms of traffic system, consumer loyalty degree at present.
Therefore the present invention is imitated based on SEM models to the direct effect of public bicycles consumer loyalty degree influence factor and indirectly
It should be analyzed, and loyalty and influence factor are evaluated, to formulate public bicycles System Development strategy and operator
Case provides effective guidance, to realize the fine-grained management of public bicycles user's request.
The content of the invention
In order to overcome the disadvantages mentioned above of prior art, the present invention proposes a kind of public bicycles based on SEM models and used
Family loyalty determines method, with reference to public bicycles characteristic and uses structural equation model, and public bicycles user is built first
The initial theory model of loyalty, proposes some latent variables and preliminary research it is assumed that according to pre- survey data on this basis
Questionnaire validity and reliability are tested and corrected, public bicycles user after revised questionnaire and amendment is obtained loyal
Initial model is spent, questionnaire data collection, questionnaire data normal distribution-test, model empirical test is secondly carried out and corrects, obtain
Final public bicycles consumer loyalty degree model, finally with reference to the practical problem of public bicycles system operation, with public
Bicycle consumer loyalty degree model carries out loyalty influence factor direct effect and indirect effect analysis and evaluated, and can be formulation
Public bicycles System Development strategy and Management plan provide effective instruct.
The technical solution adopted for the present invention to solve the technical problems is：A kind of public bicycles based on SEM models are used
Family loyalty determines method, comprises the following steps：
Step 1: the public bicycles consumer loyalty degree initial theory model based on structural equation model is built, including it is anti-
The measurement model for the relation reflected between latent variable and the therebetween structural model of relation and reflection latent variable and its aobvious variable；
Step 2: by pre- questionnaire gathered data, calculate KMO values, side reaction coefficient using questionnaire data, factor loading, tire out
Explained variance is counted, questionnaire quality is tested, questionnaire amendment and initial model amendment are then carried out according to assay, obtained
Questionnaire and revised public bicycles consumer loyalty degree model after amendment；
Step 3: by online questionnaire platform granting and reclaiming public bicycles user investigation questionnaire, questionnaire data is calculated
Polynary kurtosis value and polynary degree of bias value, normal distribution-test is carried out to the valid data of acquisition；Then each variable is calculated
Number of samples, minimum value, maximum, average and sample standard deviation；
Step 4: public bicycles consumer loyalty degree model is tested and corrected：
Using standard error, critical ratio, P values and degree of fitting index as the evaluation index of model empirical test, testing model
The significance degree of middle variable, the fitting degree that matches of model and real example data, and amendment is made to model according to test effect, directly
To model by examining；
Step 5: according to model output result, the value of utility and public bicycles for calculating each latent variable to loyalty are loyal
Spend the Satisfaction Index of variation.
Compared with prior art, the positive effect of the present invention is：For being ground at present to public bicycles consumer loyalty degree
Study carefully it is relatively simple, lack in overall aspect and influence factor cause and effect the effective analysis deeply sought with Influencing Mechanism, lack
A kind of the problems such as weary research to user behavior characteristic, satisfaction and loyalty Influencing Mechanism, it is proposed that public affairs based on SEM models
Bicycle consumer loyalty degree determines method altogether.With reference to public bicycles characteristic and structural equation model is used, built first public
The initial theory model of bicycle consumer loyalty degree, some latent variables and preliminary research are proposed on this basis it is assumed that according to
Pre- survey data is tested and corrected to questionnaire validity and reliability, questionnaire and public bicycles after amendment after being corrected
Consumer loyalty degree initial model, secondly carries out survey, questionnaire data normal distribution-test, models fitting and conspicuousness inspection
Test and correct, obtain final public bicycles consumer loyalty degree model.Advantage of this approach is that：Public bicycles are built to use
The structural equation model of family loyalty, the influence factor to public bicycles consumer loyalty degree is furtherd investigate, whole from system
The angle of body is analyzed the influence factor and cause and effect of consumer loyalty degree, on public bicycles consumer loyalty degree and its influence because
Element has carried out quantitative assessment, and proposes specific aim suggestion for system enhancement.The invention can be perfectly suitable for consumer loyalty degree
This change quantity research that can not directly observe, the use wish kept to public bicycles long-term to user, raising are public voluntarily
Car system user radix, utilization rate have very great help, and then can lift public bicycles share rate or even slow-moving traffic is shared
Rate, so as to make significant contribution in terms of urban traffic blocking is alleviated.
Brief description of the drawings
Examples of the present invention will be described by way of reference to the accompanying drawings, wherein：
Fig. 1 is the path profile citation form of structural equation model；
Fig. 2 is public bicycles consumer loyalty degree initial theory model；
Fig. 3 is public bicycles consumer loyalty degree theoretical model after amendment；
Fig. 4 is public bicycles consumer loyalty degree model path profile.
Embodiment
A kind of public bicycles consumer loyalty degree based on SEM models determines method, with reference to public bicycles characteristic and transports
With structural equation model, the initial theory model of public bicycles consumer loyalty degree is built first；For each latent variable in model
Between each variable of influence relation pair be defined and propose corresponding Research Hypothesis；Pass through the preceding validity for surveying analysis to initial questionnaire
Test and correct with reliability, questionnaire and public bicycles consumer loyalty degree initial model after amendment after being corrected,
Secondly the collection of actual data is carried out, normal distribution is carried out by the absolute value of the polynary kurtosis to questionnaire data and the polynary degree of bias
Examine, models fitting and conspicuousness are tested and corrected, obtain final public bicycles consumer loyalty degree model, most
The practical problem of public bicycles system operation is combined afterwards, based on the model set up, carries out public bicycles consumer loyalty degree
Influence factor direct effect and indirect effect analysis and evaluation, can be formulation public bicycles System Development strategy and operator
Case provides effective instruct.Following process steps are mainly included in methods described：
Step one：Public bicycles consumer loyalty degree initial theory modelling
On the basis of Customer Satisfaction Index Model (CSI), by analyzing domestic and overseas correlative development, with reference to public
Bicycle self-characteristic, proposes public bicycles consumer loyalty degree initial theory model, including reflect latent variable and therebetween relation
Structural model and reflect the measurement model of the relation between latent variable and its aobvious variable, and primarily determine that latent variable, aobvious variable
And corresponding Research Hypothesis.
Structural model describes the relation between latent variable, and its mathematical expression such as matrix equation is as follows：
η=B η+Γ ξ+ζ
In formula, η is interior raw latent variable, and ξ is external latent variable, and B represents the path coefficient between interior raw latent variable, and Γ is represented
External latent variable is to the path coefficient of interior raw latent variable, and ζ represents residual error.
Measurement model describes latent variable and aobvious relationship between variables, and its mathematical expression such as matrix equation is as follows：
X=Λxξ+δ
Y=Λyη+ε
In formula, x represents external aobvious variable, and y represents that interior life shows variable, ΛxRepresent that x shows the relation of variable and latent variable ξ, Λy
Represent that y shows the relation of variable and latent variable η, ΛxAnd ΛyIt is factor loading, δ, ε is the error in x, y measurements.
The path profile citation form of structural equation model is as shown in Figure 1.
Step 2：Pre- investigation and initial model amendment
1) pre- investigation
According to public bicycles system actual characteristic, the relation in analysis model between latent variable and aobvious variable, while simultaneous
Resident's personal information feature is turned round and look at, using 5 grades of scale forms of Likert, public bicycles consumer loyalty degree influence factor is formed
Measurable table, designs pre- questionnaire, gathered data.
2) initial model amendment
The preceding survey analysis of pre- questionnaire data is carried out from two contents of validity analysis and Reliability Analysis, questionnaire quality is entered
Performing check, is used as and is asked using KMO values, Cronbach ' s alpha coefficients (side reaction coefficient), factor loading, accumulative explained variance (%)
Roll up the evaluation index of quality inspection.Examined according to validity and credit assigned result adjust initial latent variable, and to questionnaire, just
Beginning Research Hypothesis and initial theory model are modified.
Wherein, evaluation index KMO values and the calculation formula of Cronbach's alpha coefficients：
In formula：It is the simple correlation coefficient between variable and variable,It is the partial correlation coefficient between them；N is amount
Table title mesh number, r is the average correlation coefficient between topic.
Step 3：Questionnaire data collection and analysis
By " Baidu MTC- mobile test " center "s carry out questionnaire for public bicycles user in the form of online questionnaire
Investigation, and carry out data screening and pretreatment.
Normal distribution-test is carried out to the valid data of acquisition by the polynary kurtosis of data and the absolute value of the polynary degree of bias.
By SPSS softwares to variable data being described property statistical disposition, the number of samples of each variable, minimum value, maximum are calculated
Value, average and sample standard deviation.
Step 4：The inspection and amendment of public bicycles consumer loyalty degree model
1) preliminary test of research model
Selection standard error (S.E), critical ratio (C.R), P values and degree of fitting index as model empirical test evaluation
The significance degree of relationship between variables in index, testing model, model matches fitting degree with real example data.
2) research model amendment and model are examined again
Initial test results based on model, are modified to model.Inapparent path is deleted, while defeated according to model
Go out result, add respective paths, then re-start model testing.
Step 5：The effect analysis of the influence factor of public bicycles consumer loyalty degree and evaluation
According to model output result, the influence factor to public bicycles consumer loyalty degree carries out direct effect and indirectly effect
It should analyze, and loyalty and influence factor are evaluated.
1) public bicycles consumer loyalty degree influence factor effect analysis
According to the standard routes coefficient results between each latent variable, each latent variable (i.e. influence factor) is calculated to loyalty
Value of utility.The size of direct effect is weighed with the path coefficient of causal variable to outcome variable, with logical from causal variable
Cross all intermediate variables to end at the path coefficient product of outcome variable to weigh the size of indirect effect, direct effect and indirectly
Effect sum is the gross effect between variable.
2) public bicycles consumer loyalty degree and influence element assessment
Satisfaction of the user to periphery public bicycles website quantity is evaluated using the Satisfaction Index height of variable.
User is evaluated to periphery public bicycles website number as variable index using Customer Satisfaction Index Model (CSI)
The satisfaction of amount.The CSI calculation formula of aobvious variable and latent variable are：
The formula of index of aobvious variable：
Wherein,Represent the survey data average of the aobvious variable.
The formula of index of latent variable：
In formula,The survey data average of i-th of aobvious variable in-latent variable；wi- i-th corresponding creep of aobvious variable
The nonstandardized technique factor loading coefficient of amount；The aobvious variable number that n-latent variable is included.
The inventive method is described in detail below with reference to accompanying drawing：
Fig. 1 is the path profile citation form of structural equation model, and oval symbol represents latent variable, and rectangle symbols represent aobvious change
Amount, circle symbol represents the corresponding measurement error of variable, and arrow line represents the influence relation between variable, λ represent latent variable with
Factor loading between aobvious variable,Represent the path coefficient between latent variable.Fig. 2 is the exhibition in the form of path profile by AMOS softwares
Show constructed public bicycles consumer loyalty degree initial theory model, intuitively show the latent variable included in model, show
The influence relation of variable and hypothesis.Fig. 3 is that the quality inspection evaluation index of pre- questionnaire is calculated by SPSS softwares
Analysis, the aobvious variable of adjustment, assumes preliminary research and initial theory model is modified, public bicycles user after being corrected
Loyalty theoretical model.Fig. 4 is the initial test results based on model, model is modified and to each paths of model again
Test, degree of fitting meets the significant public bicycles consumer loyalty degree model path profile of standard and path.
Hereinafter, the present invention is using certain city public bicycle system as model real example object, according to personal attribute, life cycle
Border, weather, the condition such as potential condition, public bicycles system facility level, obtain certain city public bicycle user's questionnaire number
According to public bicycles consumer loyalty degree analysis of Influential Factors technology introduction of the progress based on SEM models：
Step 1：Public bicycles consumer loyalty degree initial theory modelling
The determination of 1.1 latent variables and its relation
By studying the domestic and foreign literature on consumer loyalty degree, on the basis of Customer Satisfaction Index Model (CSI),
Use for reference SCBC models (Sweden's customer satisfaction weather table model), ACSI models (ACSI model), ECSI
The setting of latent variable in model (European Customer Satisfaction Index Model) and CCSI models (China Customer Satisfaction exponential model),
With reference to public bicycles self character, it is proposed that 9 creeps with public bicycles consumer loyalty degree associating directly or indirectly
(environment sensing, perception facility level, perceiving service quality, perceived value, riding safety are perceived amount, user expects, user's satisfaction
Degree, perceived cost and consumer loyalty degree) and 19 Research Hypothesis of relation therebetween, scale-model investigation assumes to collect such as table 1.
The scale-model investigation of table 1 is assumed to collect
The determination of 1.2 aobvious variables
With reference to and use for reference the ripe scale in domestic and international pertinent literature, public bicycles characteristic is taken into account, with can directly survey
The aobvious variable of amount substitutes the implication of 9 latent variables respectively, it is determined that the aobvious change that each latent variable is included in the structural model built
It is affected by environment to riding that amount totally 36, wherein environment sensing show as public bicycles user, perceives facility level and shows as
Public bicycles facility level, riding safety perceives the factor for showing as influenceing user to judge public bicycles security, perceives
Service quality performance is that perceived cost shows as actual use by means of imbalance problem and public bicycles service system quality is gone back
The expenditure summation felt during public bicycles, perceived value shows as user to the sense after public bicycles usage experience
Know value, user expect to show as desirable to provide service can meet demand, user satisfaction show as user using it is public from
A kind of subjective feeling that demands of individuals and expectation are met after driving, consumer loyalty degree shows as usage behavior and perception attitude
Organic unity.The corresponding each aobvious variable of latent variable is as shown in table 2 below.
1.3 model latent variables and its aobvious variable collect
Constructed model is shown in the form of path profile by AMOS softwares, is intuitively shown latent included in model
The influence relation of variable, aobvious variable and hypothesis, the public bicycles consumer loyalty degree initial theory model of formation is as shown in Figure 2.
9 latent variables, the influence relations between 36 aobvious variables, each variables are presented in initial model in figure and each variable is corresponding surveys
Measure error.Initial theory structural equation model latent variable, aobvious variable and residual error are as shown in table 2.
The initial theory structural equation model latent variable of table 2, aobvious variable and residual error
Step 2：Pre- investigation and initial model amendment
2.1 pre- investigation
Analyze the aobvious variable composition situation in 9 latent variables, with reference to the personal information of interviewee, to user's sex, the age,
Income, education degree and the main purpose gone on a journey using public bicycles and accordingly aobvious variable carry out asking that item is designed, and use
5 grades of scale forms of Likert, form the measurable table of public bicycles consumer loyalty degree influence factor, obtain initial questionnaire.
2.2 initial model amendments
Carry out questionnaire quality inspection：Analysis, including validity analysis and Reliability Analysis are surveyed before being carried out to pre- questionnaire data
Two parts are (big using KMO values (being more than 0.7), Cronbach ' s alpha coefficients (side reaction coefficient) (being more than 0.6), factor loading
In 0.5), accumulative explained variance (%) (being more than or equal to 50%) as the evaluation index of questionnaire quality inspection, by SPSS softwares
Questionnaire quality inspection evaluation index is carried out after calculating analysis, the not strong variable of correlation is rejected, for being not enough to have comprehensively
Effect reflects the latent variable of its intension, the aobvious variable of addition until meeting standard, or deletes from model the latent variable.
Questionnaire quality inspection evaluation index result sample is as shown in table 3.Wherein " physical features situation " factor loading is less than 0.5,
Represent not strong with " environment sensing " correlation；" riding safety perception ", the KMO values of " user's expectation " are below 0.7, show to be wrapped
The aobvious variable contained is also not enough to comprehensively and effectively reflect its intension.By analyzing the intension of " riding safety perception ", by " physical features feelings
Condition ", " cycling a failure-frequency " are adjusted to " riding safety perception " latent variable, and " user's expectation ", which is difficult to add again, new to be had
The aobvious variable of effect, this latent variable is deleted, therefore former 9 latent variables are adjusted to 8 (deleting " user's expectation "), while according to above-mentioned
The adjustment of questionnaire test result accordingly shows variable, realizes the amendment to questionnaire, preliminary research hypothesis and initial theory model, and
Questionnaire quality inspection is carried out, until meeting evaluation index requirement, and questionnaire and research model after amendment is formed.It is public after amendment
Bicycle consumer loyalty degree theoretical model is as shown in Figure 3 altogether.The adjustment of variable is carried out on the basis of Fig. 2 and is deleted, by " physical features
Situation ", " cycling a failure-frequency " are adjusted to " riding safety perception " latent variable, delete " user's expectation " latent variable.
The questionnaire quality inspection evaluation index result sample of table 3
Step 3：Questionnaire data collection and analysis
3.1 questionnaire data acquisitions
With questionnaire after being corrected in step 2, certain city public bicycle system is chosen as the real example pair of model
As by " Baidu MTC- mobile tests " center " carries out Online Questionnaire, and effective 341 parts of questionnaire is reclaimed altogether, structure side is met
Journey model data ideal sample amount is to study 10 times of requirement of total number of variable 33.Questionnaire data sample is as shown in table 4.
The questionnaire data sample of table 4
3.2 data multivariate normal distributions are examined
The polynary kurtosis value (Kurtosis) and polynary degree of bias value (Skew) of questionnaire data are calculated by AMOS softwares, as a result
Show that the absolute value of skewness and kurtosis of all variables, all close to 0, shows that questionnaire data relatively meets multivariate normal distributions, meet
Requirement of the structural equation model to data distribution situation.Questionnaire data multivariate normal distributions assay sample is as shown in table 5.
The questionnaire data multivariate normal distributions assay sample of table 5
The descriptive statistical analysis of 3.3 research variables
By SPSS softwares to variable data being described property statistical disposition, the number of samples (N), most of each variable is calculated
Small value (MIN), maximum (MAX), average and sample standard deviation, understand public bicycles user to itself loyalty and its shadow
The perception degree of the factor of sound.Study variable description statistic analysis result sample such as table 6.
Table 6 studies variable description statistic analysis result sample
Step 4：The inspection and amendment of public bicycles consumer loyalty degree model
The preliminary test of 4.1 research models
Referred to using standard error (S.E), critical ratio (C.R), P values and degree of fitting index as the evaluation of model empirical test
The significance degree of relationship between variables in mark, testing model, model preferably matches fitting degree with real example data, carries out diagnostic cast
The preliminary test of type.
By AMOS softwares according to the path coefficient between each latent variable of the variance and covariance of variable estimation, reflect latent variable
Between influence degree size and Orientation.Binding model empirical test evaluation index：Standard error S.E values are not negative, critical ratio
Value C.R>2, it is assumed that it is not notable, p to examine P value >=0.05<0.05 is notable, P<0.01 is highly significant, and according to model road
Footpath significance test result, deletes not notable path.Significance test result sample in model path is as shown in table 7.Can by table 5
Know, the C.R values and P values of " perceived cost ← perception facility level " and " perceived value ← riding safety is perceived " two paths are not
Meet and require, illustrate that this two paths is not notable, it should delete.It can be seen that from the fitting degree output result of model and data,
The degree of fitting result of model meets the index requests, model such as relative card side (CMIN/DF), approximate error mean square deviation (RMSEA)
Preferably, model-fitting degree assay sample is as shown in table 8 for fitting degree.
The model path significance test result sample of table 7
Note：* * represent P<0.001
The model-fitting degree assay sample of table 8
4.2 research model amendments and model are examined again
Initial test results based on model, are modified to model.Can be by deleting not notable path, adding in a model
Plus the fitting effect of the unidirectional path coefficient lift scheme between the two-way covariance relationship and variable between latent variable.By
Described in 4.1, two inapparent paths are deleted；Simultaneously according to model output result, the covariance relationship between addition variable, i.e.,
Error term e27 and e28, e13 and e17, e13 and e14 and e13 and the paths of e15 tetra- are added, model testing is then re-started.
After Modifying model, each paths of model are satisfied by test evaluation index request, the notable sexual satisfaction standard in path, while models fitting
Degree meets standard and optimized.
Factor loading value of the final model output result comprising latent variable and aobvious variable, path coefficient between each latent variable,
Model path profile and Research Hypothesis assay.Public bicycles consumer loyalty degree model path profile is as shown in Figure 4.
Step 5：The effect analysis of the influence factor of public bicycles consumer loyalty degree and evaluation
According to model output result, the influence factor of Zhongshan city's public bicycles consumer loyalty degree is carried out direct effect and
Indirect effect is analyzed, and loyalty and influence factor are evaluated.
5.1 Zhongshan city's public bicycles consumer loyalty degree influence factor effect analysis
According to the standard routes coefficient results between each latent variable in step 4, each latent variable (i.e. influence factor) is calculated right
The value of utility of loyalty, each latent variable is as shown in table 9 to the value of utility result of loyalty.Value of utility size is according to descending row
Sequence, wherein value of utility are to bear only to represent that the influence to loyalty is negative sense.It can draw, be used for Zhongshan city's public bicycles
Family, three maximum factors, perceiving service quality, perception facility level and perceived cost are influenceed on its loyalty.
Value of utility result of each latent variable of table 9 to loyalty
5.2 Zhongshan city's public bicycles consumer loyalty degrees and influence element assessment
Seen according to factor loading result between each variable of Zhongshan city's public bicycles consumer loyalty degree model and each variable
Mean data, using aobvious variable, the formula of index of latent variable, expires to Zhongshan city's public bicycles loyalty variation
Meaning degree index is calculated, and wherein questionnaire is using 5 grades of scales, and maximum is that 5 expressions are very satisfied, and minimum value is 1, is represented very not
Satisfied, Satisfaction Index carries out value according to hundred-mark system in Customer Satisfaction Index Model (CSI)：" 80-100 " is very full
Meaning, " 60-80 " is satisfaction, and " 40-60 " is general, and " 20-40 " is dissatisfied, and " 0-20 " is very dissatisfied.Satisfaction Index knot
Fruit sample such as table 10 shows.
The Satisfaction Index result sample of the Zhongshan city's public bicycles consumer loyalty degree of table 10 and variation
Claims (8)
1. a kind of public bicycles consumer loyalty degree based on SEM models determines method, it is characterised in that：Comprise the following steps：
Step 1: the public bicycles consumer loyalty degree initial theory model based on structural equation model is built, including reflection is latent
The measurement model of relation between the structural model and reflection latent variable of variable and therebetween relation and its aobvious variable；
Step 2: by pre- questionnaire gathered data, KMO values, side reaction coefficient, factor loading, accumulative solution are calculated using questionnaire data
Variance is released, questionnaire quality is tested, questionnaire amendment and initial model amendment are then carried out according to assay, obtains formal
Questionnaire and revised public bicycles consumer loyalty degree model；
Step 3: by online questionnaire platform granting and reclaiming public bicycles user investigation questionnaire, many of questionnaire data are calculated
The valid data of acquisition are carried out normal distribution-test by first kurtosis value and polynary degree of bias value；Then the sample of each variable is calculated
Number, minimum value, maximum, average and sample standard deviation；
Step 4: carrying out empirical test and amendment to public bicycles consumer loyalty degree model：
Using standard error, critical ratio, P values and degree of fitting index as the evaluation index of model empirical test, become in testing model
The significance degree of amount, the fitting degree that matches of model and real example data, and amendment is made to model according to test effect, until mould
The requirement that type passes through test evaluation index；
Step 5: according to model output result, calculating value of utility and public bicycles loyalty shadow of each latent variable to loyalty
Ring the Satisfaction Index of variable.
2. the public bicycles consumer loyalty degree according to claim 1 based on SEM models determines method, its feature exists
In：The mathematical expression matrix equation of the structural model is：
η=B η+Γ ξ+ζ
In formula, η is interior raw latent variable, and ξ is external latent variable, and B represents the path coefficient between interior raw latent variable, and Γ represents external
Latent variable is to the path coefficient of interior raw latent variable, and ζ represents residual error；
The mathematical expression matrix equation of the measurement model is：
X=Λxξ+δ
Y=Λyη+ε
In formula, x represents external aobvious variable, and y represents that interior life shows variable, ΛxRepresent that x shows the relation of variable and latent variable ξ, ΛyRepresent
Y shows the relation of variable and latent variable η, ΛxAnd ΛyIt is factor loading, δ, ε is the error in x, y measurements.
3. the public bicycles consumer loyalty degree according to claim 1 based on SEM models determines method, its feature exists
In：The KMO values and the calculation formula of side reaction coefficient are as follows：
<mrow>
<mi>K</mi>
<mi>M</mi>
<mi>O</mi>
<mo>=</mo>
<mfrac>
<mrow>
<mo>&Sigma;</mo>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>&NotEqual;</mo>
<mi>j</mi>
</mrow>
<mn>2</mn>
</msubsup>
<msubsup>
<mi>r</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mn>2</mn>
</msubsup>
</mrow>
<mrow>
<mo>&Sigma;</mo>
<msub>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>&NotEqual;</mo>
<mi>j</mi>
</mrow>
</msub>
<msubsup>
<mi>r</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mn>2</mn>
</msubsup>
<mo>+</mo>
<mo>&Sigma;</mo>
<msub>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>&NotEqual;</mo>
<mi>j</mi>
</mrow>
</msub>
<msubsup>
<mi>p</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mn>2</mn>
</msubsup>
</mrow>
</mfrac>
</mrow>
<mrow>
<mi>&alpha;</mi>
<mo>=</mo>
<mfrac>
<mrow>
<mi>n</mi>
<mi>r</mi>
</mrow>
<mrow>
<mo>(</mo>
<mi>n</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
<mi>r</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</mfrac>
</mrow>
In formula：It is the simple correlation coefficient between variable and variable,It is the partial correlation coefficient between variable and variable；N is
Scale topic number, r is the average correlation coefficient between topic.
4. the public bicycles consumer loyalty degree according to claim 1 based on SEM models determines method, its feature exists
In：It is as follows according to the method for assay progress questionnaire amendment and initial model amendment described in step 2：If assay is discontented with
The following evaluation index of foot, then reject the not strong variable of correlation；Latent variable for being not enough to effectively reflect its intension comprehensively, adds
Plus aobvious variable is until meet standard, or delete the latent variable for being difficult to add new effectively aobvious variable again；The evaluation index is
Refer to KMO values and be all higher than 0.5, accumulative explained variance more than or equal to 50% more than 0.6, factor loading more than 0.7, side reaction coefficient.
5. the public bicycles consumer loyalty degree according to claim 1 based on SEM models determines method, its feature exists
In：The number of the public bicycles user investigation questionnaire reclaimed described in step 3 needs to meet " structural equation model data ideal sample
This amount is study total number of variable 10 times ".
6. the public bicycles consumer loyalty degree according to claim 1 based on SEM models determines method, its feature exists
In：Model empirical test evaluation index is described in step 4：Standard error S.E values are not negative, critical ratio C.R>2, it is assumed that
It is not notable, p to examine P value >=0.05<0.05 is notable, P<0.01 is highly significant.
7. the public bicycles consumer loyalty degree according to claim 1 based on SEM models determines method, its feature exists
In：The method that amendment is made to model according to test effect described in step 4 is：Inapparent path is deleted, is added in a model
The unidirectional path coefficient between two-way covariance relationship and variable between latent variable, then re-starts model testing with repairing
Just.
8. the public bicycles consumer loyalty degree according to claim 1 based on SEM models determines method, its feature exists
In：The Satisfaction Index of public bicycles loyalty variation described in step 5 includes the Satisfaction Index CSI of aobvious variableAobvious variable
With the Satisfaction Index CSI of latent variableLatent variable, wherein：
In formula,The survey data average of aobvious variable is represented,Represent the survey data average of i-th of aobvious variable in latent variable；wi
Represent the nonstandardized technique factor loading coefficient of i-th of corresponding latent variable of aobvious variable；N represents the aobvious variable that latent variable is included
Number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710474219.9A CN107180160A (en) | 2017-06-21 | 2017-06-21 | Public bicycles consumer loyalty degree based on SEM models determines method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710474219.9A CN107180160A (en) | 2017-06-21 | 2017-06-21 | Public bicycles consumer loyalty degree based on SEM models determines method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107180160A true CN107180160A (en) | 2017-09-19 |
Family
ID=59844245
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710474219.9A Pending CN107180160A (en) | 2017-06-21 | 2017-06-21 | Public bicycles consumer loyalty degree based on SEM models determines method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107180160A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110414254A (en) * | 2019-08-07 | 2019-11-05 | 北京艾摩瑞策科技有限公司 | The cochain method and device thereof of user's investigation associated data |
| CN110766437A (en) * | 2019-07-15 | 2020-02-07 | 北方工业大学 | Passenger characteristic-based method for calculating loyalty of rail transit passengers |
| CN111091274A (en) * | 2019-12-03 | 2020-05-01 | 西安博达软件股份有限公司 | Dimension-based domestic medical institution online satisfaction survey method |
| CN111091411A (en) * | 2019-11-07 | 2020-05-01 | 央视市场研究股份有限公司 | Questionnaire segmentation design method |
| CN112257900A (en) * | 2020-09-23 | 2021-01-22 | 国网河北省电力有限公司经济技术研究院 | Structural equation-based power distribution network frame optimization method with distributed power supply |
| CN112308381A (en) * | 2020-10-12 | 2021-02-02 | 西安电子科技大学 | Equipment contribution degree data analysis method, system, storage medium and computer equipment |
| CN112307417A (en) * | 2020-11-26 | 2021-02-02 | 武汉筑信科技有限公司 | Structural equation model calculation technology based on formula constraint deterministic algorithm |
| CN112632765A (en) * | 2020-12-17 | 2021-04-09 | 中国兵器工业信息中心 | Combat capability assessment method combining weighted sum method and SEM method |
| CN112700283A (en) * | 2021-01-06 | 2021-04-23 | 东南大学 | Method for improving shared automobile utilization rate based on latent variable |
| CN112988732A (en) * | 2021-04-14 | 2021-06-18 | 湖南工程学院 | Method for processing abnormal value in observation data |
| CN113110990A (en) * | 2021-03-25 | 2021-07-13 | 浙江工业大学 | Virtual simulation software user experience evaluation method based on structural equation model |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106469177A (en) * | 2015-08-21 | 2017-03-01 | 中国传媒大学 | Internet film user satisfaction analysis method based on structural equation model and system |
-
2017
- 2017-06-21 CN CN201710474219.9A patent/CN107180160A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106469177A (en) * | 2015-08-21 | 2017-03-01 | 中国传媒大学 | Internet film user satisfaction analysis method based on structural equation model and system |
Non-Patent Citations (3)
| Title |
|---|
| SHANG-YU CHEN: "Using the sustainable modified TAM and TPB to analyze the effects of perceived green value on loyalty to a public bike system", 《TRANSPORTATION RESEARCH PART A》 * |
| 钱佳 等: "城市居民使用市内公共自行车的满意度影响因素分析——以苏州市为例", 《地理研究》 * |
| 陈坚 等: "出行方式选择行为的SEM-Logit整合模型", 《华南理工大学学报（自然科学版）》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110766437A (en) * | 2019-07-15 | 2020-02-07 | 北方工业大学 | Passenger characteristic-based method for calculating loyalty of rail transit passengers |
| CN110414254A (en) * | 2019-08-07 | 2019-11-05 | 北京艾摩瑞策科技有限公司 | The cochain method and device thereof of user's investigation associated data |
| CN111091411A (en) * | 2019-11-07 | 2020-05-01 | 央视市场研究股份有限公司 | Questionnaire segmentation design method |
| CN111091411B (en) * | 2019-11-07 | 2023-12-22 | 央视市场研究股份有限公司 | Questionnaire segmentation design method |
| CN111091274A (en) * | 2019-12-03 | 2020-05-01 | 西安博达软件股份有限公司 | Dimension-based domestic medical institution online satisfaction survey method |
| CN112257900A (en) * | 2020-09-23 | 2021-01-22 | 国网河北省电力有限公司经济技术研究院 | Structural equation-based power distribution network frame optimization method with distributed power supply |
| CN112308381B (en) * | 2020-10-12 | 2023-06-20 | 西安电子科技大学 | Equipment contribution data analysis method, system, storage medium and computer device |
| CN112308381A (en) * | 2020-10-12 | 2021-02-02 | 西安电子科技大学 | Equipment contribution degree data analysis method, system, storage medium and computer equipment |
| CN112307417A (en) * | 2020-11-26 | 2021-02-02 | 武汉筑信科技有限公司 | Structural equation model calculation technology based on formula constraint deterministic algorithm |
| CN112632765B (en) * | 2020-12-17 | 2023-09-22 | 中国兵器工业信息中心 | Combat capability assessment method combining weighting method and SEM method |
| CN112632765A (en) * | 2020-12-17 | 2021-04-09 | 中国兵器工业信息中心 | Combat capability assessment method combining weighted sum method and SEM method |
| CN112700283A (en) * | 2021-01-06 | 2021-04-23 | 东南大学 | Method for improving shared automobile utilization rate based on latent variable |
| CN113110990A (en) * | 2021-03-25 | 2021-07-13 | 浙江工业大学 | Virtual simulation software user experience evaluation method based on structural equation model |
| CN112988732A (en) * | 2021-04-14 | 2021-06-18 | 湖南工程学院 | Method for processing abnormal value in observation data |
| CN112988732B (en) * | 2021-04-14 | 2023-10-20 | 湖南工程学院 | Method for processing abnormal value in observed data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107180160A (en) | Public bicycles consumer loyalty degree based on SEM models determines method | |
| Haas et al. | Compensating for non-homogeneity in decision-making units in data envelopment analysis | |
| Wang et al. | A hybrid Data Quality Indicator and statistical method for improving uncertainty analysis in LCA of complex system–application to the whole-building embodied energy analysis | |
| CN103412918B (en) | A kind of service trust degree appraisal procedure based on service quality and reputation | |
| Amador et al. | Preference heterogeneity and willingness to pay for travel time savings | |
| Gim | A meta-analysis of the relationship between density and travel behavior | |
| CN101916334B (en) | A kind of skin Forecasting Methodology and prognoses system thereof | |
| CN103235877B (en) | Robot control software's module partition method | |
| CN109243178A (en) | Town way Traffic Safety Analysis and evaluation method under the conditions of a kind of bad climate | |
| CN101853290A (en) | Meteorological service performance evaluation method based on geographical information system (GIS) | |
| CN102467684A (en) | System and method for evaluating comprehensive performance based on improved radar chart | |
| CN102968551B (en) | A kind of electric automobile operation characteristic modeling and analysis methods | |
| CN107273605B (en) | Actually measured axle load spectrum determination method based on multiple classifier system | |
| Abreu et al. | Ordinal logistic regression in epidemiological studies | |
| Wang et al. | A multi‐criterion experimental comparison of three multi‐attribute weight measurement methods | |
| Wolman | Measurement and meaningfulness in conservation science | |
| CN106056274A (en) | Power construction main body benefit analysis method based on PCA-DEA two-dimensional comprehensive evaluation model | |
| Batarce et al. | Urban travel demand model with endogenous congestion | |
| CN107256445A (en) | Analysis method, the device and system of production environment security | |
| CN101976222A (en) | Framework-based real-time embedded software testability measuring method | |
| Guo et al. | Predicting the duration of a general contracting industrial project based on the residual modified model | |
| CN106127419A (en) | A kind of construction project decision system based on group decision and decision method | |
| CN109657967A (en) | A kind of confirmation method and system of Transmission Expansion Planning in Electric evaluating indexesto scheme weight | |
| Xie | Measuring regional variation in sex preference in China: A cautionary note | |
| Bhat et al. | A multivariate hurdle count data model with an endogenous multiple discrete–continuous selection system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170919 |
|
| RJ01 | Rejection of invention patent application after publication |