时间：2022-10-28 12:28:57
Abstract: The progress and development of the society promotes service industry to grow, and service-oriented enterprises are emerging. However, the lagging service theory makes the development of service economy lack of the corresponding theoretical support. Based on analyzing the sources of corporate profits in service economy era, the paper uses data mining technique to apply customer classification to measurement methods of service capability. The change of service capability not only makes the enterprises position the problem clients accurately and rapidly, but also can timely adjust business strategies, which can increase profits.
Key words: service capability, measurement methods, data mining technique
1 Introduction
The theory that service is the derivative of manufacturing industry was proposed by Rathmell in 1974. In the academic field, the research on service management falls behind that on manufacturing management, the reason for which is false definition on service. The existing research on service is based on a single industry, which hinders the generation of cross-industry thought. The research on service introduces new conceptual framework and analysis methods for service problems. Christopher H.Lovelock pointed out that deeper understanding of service sales not only makes marketing skills of traditional service industry more exquisite, but also has a significant influence on management practice of using service to drive manufacturing industry. Meanwhile, most scholars researching service admit that they lean lots of knowledge which can be applied to service from production and human resources.
It is the objective of each enterprise to improve service capability and meet the requirements of customers. But different customers have different requirements, and it is impossible for the enterprises to meet the service requirements of customers. So it is very important to improve service capability of enterprises, which has great significance for long-term stable development of enterprises.
2 Brief Description of Data Mining
Data mining means the process of extracting implicit and useful information and knowledge from lots of incomplete, noisy, fuzzy and random data. Data mining is a promising and thriving subject of database system and new database application, and is one of the most active branches of database research, development and application. It doesn’t dissatisfied with making simple query on data, but wants to find more useful knowledge from massive data by using effective data processing methods. Therefore, the process of data mining is called the process of knowledge discovery. The original data is regarded as the source of knowledge, just like mining from ore. Original data can be structured such as data in relational data base, and some can be semi-structured such as text, graphic and image data. And original data even can be heterogeneous data distributed on the network. The methods discovering knowledge can be mathematical or non-mathematical and deductive or inductive. The discovered knowledge can be used for information management, query optimization, decision support, process control and maintenance of data.
From the perspective knowledge discovery, data mining can be seen as one basic step of knowledge discovery in database. The process of knowledge discovery is as follows.
(1) Data cleansing: undoing noise or inconsistent data.
(2) Data integration: various data sources can be combined together.
(3) Data noise: retrieving and analyzing related data in database.
(4) Data conversion: the data is converted or unified into the form fitting mining such as summary or aggregation.
(5) Data mining: intelligent methods are used to extract data pattern.
(6) Mode assessment: interest is used to measure and identify really interesting mode of knowledge.
(7) Knowledge representation: using visualization and knowledge representation technique to provide mined knowledge for the user.
Data mining function is used for the mode type which needs to be sought in data mining task. Data mining tasks are generally divided into description and prediction. Descriptive tasks describe the basic features of data in database. And predictive mining tasks make deduction on the existing data for prediction. Data mining system should be able to excavate many types of modes to meet the requirements of different users and different applications.
3 Analysis of Automobile Maintenance Service Industry
Automobile maintenance service uses the potential and real users as service objects, and all business activities of enterprises focus on customers. With more and more evident buyers’ features of buyers’ in automobile market, homogeneous competition of automobile market is more and more fierce, and car users or consumers have more and more choices. Automobile service enterprises must start from the requirements of customers to determine the operation objective and idea, which can finally realize corporate profits. The participation of customers is higher in operation process of automobile maintenance enterprises, so customer satisfaction becomes the important indicator of operation and management level of enterprises. For example, car manufacturers make examination on the members of sales and service network including automobile operators and special maintenance stations every year, and the survey on customer satisfaction is an important content. It is an important operation and management task for automobile maintenance service enterprises to improve customer satisfaction.
In order to realize great economic benefit, automobile maintenance service enterprises must expand market share of service brand. And market share can be represented by relative market share and absolute market share. It reflects the position of enterprises in market. It may express the proportion of maintenance income of the enterprise to the total maintenance income of the market. The improvement of market share depends on customer satisfaction in the market with increasingly evident homogenization of products. Therefore, one of the most important task for all automobile maintenance enterprises is to improve customer satisfaction. And automobile maintenance service management must be adjusted with improving customer satisfaction. The management task of automobile service enterprises is to fully use internal and external available resources, make plan, organization, direction, coordination and control on service activities, improve customer satisfaction and customer loyalty, and improve market share of service of the enterprise, which can achieve the objective of realizing the best economic benefit of enterprises.
The management strategy of automobile maintenance enterprises must focus on comprehensive customer satisfaction. The key to enterprise operation is to win market and customers. And the enterprises with total comprehensive satisfaction and market share can win in the fierce competition.
4 Application of Data Mining Technique for Automobile Maintenance Service Capability Measurement
4.1 Application conditions of data mining technique for automobile maintenance service capability measurement
As a tool, effective use of data mining is based on massive and reliable data. Does the enterprise service capability have the condition?
The answer is definitely yes. With the increase of information level and introduction of management information systems, the enterprises have had complete information system. In daily operating management, the enterprises not only can get massive data about products, but also can achieve various data and information of customers. The essential data of enterprise service information can fully choose, preprocess and convert the data relating to service, which prepares for data mining. And it not only lays foundation for the research, but also makes the research possible.
Applying data mining technique needs to know the following problems. The first problem is that what problems are solved by using data mining. The second problem is data preparation for data mining. And the third problem is analytic algorithms of data mining. The paper researches application of data mining to enterprise service capability measurement, and uses data mining to solve the problem of customer segmentation of service capability. Enterprise service capability is assigned by customers. All consumers are the customers of enterprises, and are the reflection of service capability, so the final objective of service capability research is customer satisfaction. Different customers have different requirements. And the customers can be divided into major clients and secondary clients. So it is very important to make necessary customer management besides creating customer satisfaction. Data mining generally predicts the prospect of market in marketing field, and predicts long-term customer set in bank and insurance industry. The enterprise service capability in the paper introduces the concept of customers, implements effective researches on service capability, classifies the customers by data mining, analyzes and solves the reasons causing the change of service capability, and decides marketing strategy such as how to keep the relationship with loyal customers and improve service capability of enterprises, and how to make general customers become loyal customers. The predictive analysis method and aggregation method of data mining has been widely applied in many fields. The data mining for customer segmentation can use the above methods.
4.2 Data mining models used in the paper
The evaluation and measurement methods of service capability demand to classify the customers in enterprise operation. It has great significance for enterprise development strategy to master composition and types of customers. Data mining technique can extract consumption information records form databases of enterprises, and finds the relation and mode, which objectively reflects the composition of enterprise customers. The data mining techniques used in the paper are classification and prediction.
The following is an introduction on mining process of data mining technique which can the problems.
(1) General process of data mining
① Preprocessing data. Collecting and purifying information from data source and storing it. It is generally stored in data warehouse.
② Model search. Data mining tools are used to search models in data. The process can be executed by system automatically, which means that the original data can be searched from bottom to up to discover the relationship between them. And user interaction process can be used, which means that analysts ask questions initiatively to find and verify the hypothesis.
③ Results analysis. The search process of data mining needs to be repeated, the reason for which is that some new problems or requirements demand to make exquisite query after analysts evaluate output results. And the final results reprot is produced.
④ Knowledge assimilation. The results report is explained. The corresponding measures are taken based on the results, which is a manual process.
(2) Introduction of mode
Using the mode not only can find customer types in organizations, but also can judge the type of a consumer.
① The sample set to be classified is established on data of data warehouse. And the objects to be classified are called samples such as h1, h2, …, hn, and H={h1, h2, …, hn} is sample set. In order to classify the samples rationally, the properties should be quantified. And the quantified properties are called sample indicators. If there are m indicators, m-dimension vector can be used to describe samples,
hi=(hi1, hi2, …. , hin)i=(1, 2, …, n)
In actual data, the acquired data is generally not the number of closed interval[0, 1], so the original data should be standardized, and the mean value should be solved firstly. If there are n samples in sample set, the indicator k of the sample can achieve n data. In h'1k, h'2k, …, h'nk, h'nk means the data of the i sample achieved in the i sample. And the mean value is calculated according to formula 2.
=(h'1k+h'2k++h'nk)/n=/n k=1, 2, …, m (1)
Then, the standard deviation Sk of the original data is figured out according to formula 2.
And the standadized value h"ik of each data is figured out according to formula 3.
And we can get the standardized data h"ik. If it is not in the closed interval [0, 1], the following extremum standardization formula 4 can be used.
and means the maximum and minimum of h"1k, h"2k, …, h"nk.
② Establishing patterns’ similarity relationship R
R means similar matrix, and the general form is as follows.
There are many methods of calculating rij, and the maximum and minimum method is used here.
③ Cluster analysis
The maximum tree method is used, which means that a special diagram is constructed, and the classified objects are vertexes. When 0, the vertex i and vertex j can connect one side. And the practice is that i of vertex set is firstly drawn, then the sides are connected until all vertexes are connected, which can get a maximum tree or empowerment tree. Each side can empower a weight, rij. There are different connection methods, so the maximum tree is not unique. Then, the maximum takes λinterception, which means to remove the sides of rij
(3)Prediction
① Figuring out the average index of each mode
The average indexes of the achieved modes are figured out by using the following formula.
i=1, 2, …, s j=1, 2, …, m (7)
In the formula, s means the total number of modes, and p means the total number of records of the mode.
② The sample X(X1, X2,…, Xn) to be predicted is n fuzzy subsets of the sample on domain X. It is compared with the classified mode in data warehouse, and the approach degree between them is solved.
(8)
( and means the inner product and outer product of fuzzy operation. )
According to the principle of selecting near,
(9)
The mode of the samples is judged, and the results of the mode is predicted.
5 Conclusion
Based on analyzing the background that economic service arrives, the paper makes detailed analysis on business-oriented development. And the paper makes deep analysis on profit source of matured industries and service-oriented enterprises, and gets the conclusion that service is one of the main profits source of enterprises. And the characteristics of service are analyzed. The paper discusses the characteristic that service is a new commodity and expounds the contribution of loyalty customers to corporate profits. And the paper demans to implement specific measures to improve service, maintain loyalty customers, develop customer satisfaction and improve the service capability of enterprises, which can increase corporate profits. And the paper deeply analyzes the important significance of customer satisfaction for service enterprises.
Reference
[1] Huang Haixia, Xu Xianglin, Research and analysis on student achievement based on C4.5 algorithm [J], Computer Knowledge and Technology, 2011,14:3405-3409.
[2]Luo Liangping, Application of data mining technique to e-commerce field [J], Modern Marketing (Academic Edition), 2011,07:259-261.
[3] Yu Yonghong, Xiang Xiaojun, Gao Yang, Shang Lin, Yang Yubin, Research on service-oriented cloud data mining engine [J], Computer Science and Exploration, 2012,01:46-57.
[4] Xiao Jianqiong, Song Guoqin, Research on mining based on interest-correlation rule and application to recommendation course system [J], Intelligent Computer and Application, 2012,05:73-74+77.
[5] Xiao Wanrong, Shi Shaoting, Zhang Long, Zeng Shuoxun, Research on establishment of knowledge service platform based on industry――taking construction of knowledge service platform of pillar industry in Gansu [J], Gansu Science and Technology,2012,22:45-48.
[6] Zhou Shengchen, Shi Xunzhi, Qu Wenting, Shi Yingzi, Sun Yuncheng, Research on prediction of stock market time sequence based on Microblogging search and SVM [J], Computer and Modernization, 2013,04:22-26..