Research On Dynamic Monitoring Method Of Intermittent Service Process Quality Based On Data Drive

Discontinuous service is one of the most important service forms in modern service industry.It is characterized by discontinuous,transient,fluctuating and interactive complexity,and widely exists in many service fields such as accommodation,tourism,medical and health care,and housekeeping.With the vertical and horizontal expansion of intermittent service business,the number of intermittent service business system is increasing,and the core business is more and more dependent on the dynamic operation of monitoring system.However,the scope and granularity of the existing basic monitoring system cannot meet the modern and digital needs of the comprehensive systematic monitoring and management of discontinuous service process quality due to the lack of technology,means and methods.The integration of "digital world" and "physical world" is deepening day by day,promoting the transformation of intermittent service process quality monitoring from decentralized and static "experience-driven" mode to centralized and dynamic "data-driven" modern mode.However,the "data-driven" dynamic monitoring of discontinuous service process quality still faces severe challenges in monitoring logic,multi-source data integration,monitoring technology and methods.Therefore this article focuses on the dynamic monitoring method of data driven intermittent service process quality based on the content of five aspects,by taking how to form a data-driven intermittent global linkage service process quality control and command system of strategic questions as a starting point,how to carry out multi-perspective,highdimensional,evidence-based and verifiable dynamic monitoring of intermittent service process quality through the method of integration of modern information technology and management science as the basic point.(1)5W1H Dynamic monitoring framework of discontinuous service process quality based on TRIZ: combining TRIZ theory with management innovation,a data-driven logical framework of dynamic monitoring of discontinuous service process quality is constructed through object field analysis.Then 5W1 H analysis method is used to solve the direction,structure and execution problems of dynamic monitoring logic framework.And on this basis,a rolling dynamic monitoring system of "event-data-quality" is constructed.(2)The 7W3 H discontinuous service process quality multi-source data integration method based on THE E-R model: First of all,the multi-source heterogeneous service process quality problem records data and its expanded substance content are extracted based on the ER model,and the entity,feature,relationship and attribute are matched.After that,7W3 H is used to extract,associate,transform and merge data.The construction of this method have built the bridge between data and quality monitoring business,forming a data-driven data foundation.(3)SN-BTM based four-level intermittent service process quality problem identification and dynamic monitoring method:(1)Data noises are removed in the text layer;(2)At the semantic network layer,the hidden entities,attributes,relationships,topics or frameworks in the intermittent service process quality problems are revealed by network graph,density,centrality and subgroup network analysis methods.(3)In the quality problem identification layer,a service process quality problem topic identification method combining pseudo documents and BTM is constructed.The pseudo documents on the data set are greatly improved in time-effectiveness,F-measure and other aspects,and the topic cohesion of THE BTM topic model is guaranteed.(4)According to the time characteristics of discontinuous service,dynamic monitoring of subtopics is carried out.According to the logic from detailed analysis to overall induction and then to dynamic monitoring,the method solves the complicated problems of the quality of intermittent service process and can not be systematically monitored,and realizes the deeper interpretation of internal logic,thus forming a monitoring network of different forces in time,space and subject.(4)The distribution difference law of intermittent service process quality problems based on R-Q factor analysis and the monitoring method of cause law on grounds of ARM-RCA.(1)according to the customer and intermittent service enterprises under different angles of attribute data related service process quality problem of R-Q analysis model is set up contingency table,through the convergence of R and Q factor analysis measure model as well as the factors important degree,and the relationship between different variables,reveal problems distribution difference law under different perspective;(2)FP-Growth algorithm was used to excavate the hidden correlation between the attributes of intermittent service process quality problems,and the root cause analysis method was used to systematize them,revealing the causes of problems from different perspectives.The construction of the above methods tackles the problem of unclear formation and evolution logic of intermittent service process quality problems,realizes the dynamic traceability of variables between distribution and cause,constructs parallel two-way monitoring,investigates the monitoring differences of different perspectives,and forms the monitoring network of different forces.(5)Dynamic monitoring method for emerging quality problems of intermittent service process quality.(1)A new service process quality problem classification method based on Word2 vec and machine learning is constructed.In this method,the 64 dimensional text vector dataset of intermittent service process quality was constructed by Word2 vec model,and the mapping relationship between text vector data and quality problem types was learned by integrating multiple machine learning algorithms according to the algorithm and data characteristics.The optimal algorithm model was selected by training,and the new quality problem types were predicted dynamically.(2)A multi-dimensional attribute hierarchical prediction method for new service process quality problem based on integrated machine learning is constructed.This method integrates multiple machine learning algorithms to learn,train and optimize the mapping relationship of text vector data and multidimensional attributes,selects and integrates the optimal algorithm model one by one,and dynamically predicts the multidimensional attributes of new quality problems step by step.(3)A dynamic recommendation method of key service remedy measures for new service process quality problems was constructed based on multiple regression and BP neural network.This method constructs the functional relationship between satisfaction degree and service remedy measures through multiple regression analysis,and determines the key service remedy measures by comparing partial correlation coefficient and standardization coefficient.On this basis,key service remediation measures are set as labels,text sentence vectors are represented as features,and high-level features are directly obtained from the data,which are transformed into BP neural network dynamic recommendation method based on text features: The BP neural network was constructed and activated.According to the gradient descent method,the FIT()training model was used for forward training and reverse adjustment,and the weight and bias were adjusted and calculated by multiple iterations.When the error reached the preset accuracy or the learning times were greater than the maximum number of designed times,the training was finished.Finally,the optimal BP neural network model is invoked and the new key service remedy measures are recommended dynamically.The construction of the above methods verifies the difference between the algorithm models,solves the theoretical and methodological basis of monitoring from the "happened past" to the "coming future",and realizes the data-driven closed-loop system of the intermittent service process quality dynamic monitoring.In order to ensure the uniformity and scientificity of the monitoring system,the service process quality data of the accommodation industry is used separately to verify the validity and reliability of the above methods and algorithms from the design and implementation of the method system to the application and evaluation.Therefore,this study achieves the purpose of data-driven global linkage monitoring and command system foundation construction,and innovates in the research perspective,research paradigm and monitoring method construction,providing a theoretical and methodological basis for the scientific management of service process quality in the intermittent service industry.