Назад в библиотеку

International Journal of Human and Social Sciences 5:11 2010

Knowledge Discovery Techniques for Talent Forecasting in

Human Resource Application

Hamidah Jantan, Abdul Razak Hamdan, and Zulaiha Ali Othman

AbstractHuman Resource (HR) applications can be used to provide fair and consistent decisions, and to improve the effectiveness of decision making processes. Besides that, among the challenge for HR professionals is to manage organization talents, especially to ensure the right person for the right job at the right time. For that reason, in this article, we attempt to describe the potential to implement one of the talent management tasks i.e. identifying existing talent by predicting their performance as one of HR application for talent management. This study suggests the potential HR system architecture for talent forecasting by using past experience knowledge known as Knowledge Discovery in Database (KDD) or Data Mining. This article consists of three main parts; the first part deals with the overview of HR applications, the prediction techniques and application, the general view of Data mining and the basic concept of talent management in HRM. The second part is to understand the use of Data Mining technique in order to solve one of the talent management tasks, and the third part is to propose the potential HR system architecture for talent forecasting.

KeywordsHR Application, Knowledge Discovery in

Database (KDD), Talent Forecasting.

I. INTRODUCTION

RECENTLY, research in Human Resource (HR) applications that are embedded with Artificial Intelligent (AI) techniques can solve unstructured and indistinct decision making problems. These applications can help decision makers to solve inconsistent, inaccurate, inequality and unpredicted decisions. In the advancement of AI technology, there are many techniques that can be used to upgrade the capabilities of HR application. Knowledge Discovery in Database (KDD) or Data Mining is one of AI technology that has been developed for exploration and analysis in large quantities of data to discover meaningful patterns and rules. In actual fact, such data in HR data can provide a rich resource for knowledge discovery and decision support tools. So far, the techniques and application of Data Mining have not attracted much attention in Human Resource Management (HRM) field[1]. In this study, we attempt to use this approach to handle the issue in managing talent i.e. to identify existing talent by predicting their performance using the past experience

knowledge.

H. Jantan is with Faculty of Information Technology and Quantitative Science, UiTM Terengganu, 23000 Dungun, Terengganu (e-mail: hamidahjtn@ tganu.uitm.edu.my).

A. R. Hamdan and Z. A. Othman are with Faculty of Information Science and Technology UKM, 43600 Bangi, Selangor (e-mail:arh@ftsm.ukm.my).

Basically, HRM is a comprehensive set of managerial activities and tasks concerned with developing and maintaining a competent workforce-human resource. HRM aims to facilitate organizational competitiveness; enhance productivity and quality; promote individual growth and development; and complying with legal and social obligation[2]. Besides that, in any organizations, they need to compete effectively in term of cost, quality, service or innovation. All these depend on having enough right people, with the right skills, deployed in the appropriate locations at appropriate points in time.

Nowadays, in HRM field, among the challenges of HR professionals are managing talent, especially to ensure the right person for the right job at the right time. These tasks involve a lot of managerial decision, and which it is sometime very uncertain and difficult to make the best decisions. In reality, current HR decision practices depend on various factors such as human experience, knowledge, preference and judgment. These factors can cause inconsistent, inaccurate, inequality and unforeseen decisions. As a result, especially in promoting individual growth and development, this situation can often make people feel injustice. Besides that, in future, this can influence organization productivity. In talent management, to identify the existing talent is one of the top HR management challenges[3]. This challenge can be manage by using Data Mining technique in order to predict the suitable talent based on their performance. For that reason, this study aims to suggest HR system architecture using Data Mining technique for talent performance prediction. This application can be use to help managers to allocate the right person at the appropriate locations at the right time.

The rest of this paper is organized as follows. The second section describes the background of HR application; and the prediction applications and intelligent techniques used. The third section explores the overview of Data Mining and how Data Mining techniques apply in HR application. The basic concept of talent management is presented in Section 4. Section 5 describes the Data Mining techniques for talent management. Then, section 6 discusses the potential HR system architecture for talent forecasting. Finally, the paper ends with Section 7 where the concluding remarks and future research directions are identified.

694

International Journal of Human and Social Sciences 5:11 2010

II.HUMAN RESOURCE APPLICATION AND PREDICTION

A. Human Resource(HR) Application

Nowadays, HR has been linked to increased productivity, good customer service, greater profitability and overall organizational survival. To reach such link, management must not only face current issues of human resource management but also deal with future challenges to human resource management effectively[4]. Human Resource Management (HRM) tasks involved a lot of managerial decisions, according to DeCenZo[5], HR professionals need to focus the goal for each of HR activities: Staffing to locate and secure competent employees; training and development to adapt competent workers to the organization and help them obtain up-to date skill, knowledge and abilities; motivation to provide competent and adapted employees who have up-to date skill, knowledge and abilities with an environment that encourages them to exert high energy level and maintenance to help competence and adapted employees who have up-to date skill, knowledge and abilities and exerting high level energy level to maintain their commitment and loyalty to the organization.

The challenges of HRM professionals are health, HIV/AIDS, managing talent, employee rewards, retention, training and development, technology, tribalism, nepotism and corruption. On the other hand, among the major potential prospects for HRM is technology selection and implementation[6]. The benefits of technology applications in HRM are to easily deliver information from the top to bottom workers in an organization, easily to communicate with employees and it is easier for HR professionals to formulate managerial decisions. For these reasons, HR decision application can be used to achieve the HR goals in any type of decision making tasks. The potentials of HR decision applications are increased productivity, consistent performance and institutionalized expertise which are among the system capabilities embedded into specific programs[7].

In this study, we found that research in HR Decision Systems basically focuses only for the specific HRM domains such as in personnel selection, training, scheduling and job performance. Besides that, most of the HR decision applications are using expert system or Knowledge-based system (KBS) approaches such as for personnel selection and training[8]. The commercial emergence of Knowledge- based information technology systems (KBS) represents a tremendous opportunity to enhance the practice of human resource management[9]. The KBS benefits are more permanent, easier to duplicate, less expensive and automatically documented. Besides that, the limitations of KBS systems are difficult to capture informal knowledge; knowledge has not been documented and difficult to verbalize. Techniques used to verify and validate conventional systems are considered to be inadequate and KBS-specific methods are still immature. For that reasons, new HR decision system research are using other intelligent approaches such as for personnel selection, they use Data Mining approach [10, 11] and Neural Network approach [12, 13]. From this study, we found that not many research done in HR decision systems area. Besides, the problem

domains that they tried to solve are also limited to the specific domains. In information technology era, HR applications are used as a tool to support human resource managers in their decision making process.

B. Intelligent Techniques in HR Application

HR application is a part of Decision Support System (DSS) which is used to support decision making process. Nowadays, the advancement of Artificial Intelligent technologies has contributes to new DSS application than known as Intelligent Decision Support System (IDSS). IDSS is developed to help decision makers during different phases of decision making by integrating modeling tools and human knowledge. IDSSs are tools for helping decision making process where uncertainty or incomplete information exists and where decisions involving risk must be made using human judgement and preferences. Basically, an IDSS is also known as a possible theoretical model of incorporation by adapting an existing DSS system to execute in an Expert System style, such adapted systems are considered by many DSS researchers to be IDSS with the focus on the functioning of ‘man and machine’ together. Most researchers agree that the purpose of IDSSs is to support the solution of a non-structured management and enable knowledge processing with communication capabilities[14, 15].

Besides that, IDSS can incorporate specific domain knowledge and perform some types of intelligent behaviors, such as learning and reasoning, in order to support decision- making processes[15, 16]. The need to incorporate domain knowledge and intelligent capabilities in decision support system has been identified in various forms and models by many researchers. Incorporating knowledge component (through case base, rule base, knowledge acquisition subsystem or domain models) and intelligent component (through an intelligent advisory system, intelligent supervisor or model solver) can produce the intelligent applications. Intelligent behaviors are presented by an intelligent system related to the abilities of gathering and incorporating domain knowledge, learning from the acquired knowledge, reasoning about such knowledge and when enquired, being able to issue recommendations and justify outcomes.

IDSS as its name implied, is used to support decision making and not intended to replace the decision maker’s task. In addition, IDSS works under an assumption that the decision maker is familiar with the problem to be solved. IDSS gives full control to the user regarding information acquisition, evaluation and making the final decision. Nowadays, there are quite a number of computer applications that apply intelligent techniques and use DSS concepts and components. However, some researchers claim that it is an essential of DSS which uses the conventional name known as IDSS and others classify it as a member of intelligent system. In this case, the application’s name is given based on the intelligent techniques that they use, such as expert system which is uses rule based system, knowledge based system (KBS), fuzzy sets, Neural Network for reasoning and learning capabilities. Most of the IDSS applications are specifically used for problem domain in that

695

International Journal of Human and Social Sciences 5:11 2010

particular area. IDSS has consolidated the intelligent behaviors in its inference engine component. These abilities are used to support the decision making processes. In our study, we found that there are various types of intelligent techniques that are applied in IDSS applications such as knowledge base system data warehouse, fuzzy set theory, ANN, rough set classifier, multi agent and etc.[8] From the literature, we found that most applications used knowledge- based system and they agree with the advantage of using this technique. Besides it is easier to understand and implement, the KBS using rule-based and it also supports the reasoning capabilities. Research and system developments in this field increase year by year with new ideas and approaches. In that case, some of the IDSS applications use hybrid techniques to advance the system capabilities by integrating more than one intelligent technique in their application. This approach makes the IDSS applications more capable to do learning and reasoning processes. Most researchers agree that these intelligent techniques are more suitable for learning and reasoning activities. In this study, we found that most of the HR applications use expert system and some of them use data mining approaches. For these reasons, to solve problems in expert system approach, hybrid intelligent techniques can be the most effective when they are embedded with the HR application. In HRM, there are several tasks that can be solved using this approach, for examples, selecting new employees, matching people to jobs, planning careers paths, planning training needs for new and old employee, predicting current employee performance, predicting future employee and etc. In this study, we focus our study on how to embed the HR application with Knowledge Discovering in Database (KDD) approach especially for prediction purposes. HR applications can produce precise decisions when incorporated with intelligent capabilities. From the literature study, some intelligent techniques that use in HR applications are listed in Table I.

TABLE I

INTELLIGENT TECHNIQUES USED IN HR APPLICATIONS

Intelligent

HR DSS Applications

Techniques

 

 

 

Knowledge-

Web-based Training Expert System [17], Job

based System/

applicants selection [18] Personnel selection

Expert System

[7],E-training [19]

 

 

Data Mining

Job Attitudes [20], Recruit and Retain Talents

 

[21],Personnel selection [12] & [10], Project

 

Assignment [22]

 

 

Software agent

Meeting Scheduler [23]

 

 

Fuzzy set/logic

Prioritization of Human Capital [24]

 

 

Artificial Neural

Personnel selection [12]

Network

 

 

 

most of the current HR applications use other intelligent techniques to advance the capabilities of the applications. In this study, we have found researches that use AI techniques in HR field are very limited. Besides, the problem domains that they try to solve are also limited to the specific problem domains especially in personnel selection and training.

Researches and system development in this field increase year by year with new ideas and approaches. In that case, some HR applications use hybrid intelligent techniques to advance the capabilities of the existing techniques. They integrate more than one intelligent technique in their application to be more capable in explaining, learning, reasoning and forecasting processes, which can produce quite similar decisions as human decisions. Nowadays, researches have shown an increase interest in predicting human performance [22, 25, 26]. However, there has been little discussion about prediction employee’s performance which relates to human resource problem domains[22]. Basically, prediction applications by using past experience knowledge or Data Mining can be used for several tasks in HR activities such as selecting new employees, matching people to jobs, planning careers paths, planning training needs for employees, predicting employee performance, predicting future employees and etc. All these need a lot of attention and efforts, from both academicians and practitioners to explore and analyze the existing data and to discover useful knowledge.

C. Prediction: Applications and Techniques

Prediction is a process to gain knowledge about uncertain events that are important to present decisions[27]. Besides that, prediction methodology can be categorized into two approaches; statistical and intelligent techniques. In this study, we focus on intelligent techniques approaches. Some of intelligent techniques used in prediction application are listed in Table II.

TABLE II

PREDICTION TECHNIQUES AND APPLICATION

Techniques used

Applications

 

 

 

Electricity energy consumption

Decision Tree

[28]

Medicine [29]

 

 

Accident frequency [30]

 

Electricity energy consumption

 

[28]

 

Country investment risk [31]

Artificial Neural Network

Stock market returns [32]

Medicine [29]

 

 

Interest rates [33]

 

Disease [34]

 

Corporate failure[35]

 

 

Bayesian Belief Networks

Student performance [26] &[25]

(BBN)

 

Fuzzy Clustering

Newspaper demand [36]

 

 

Basically, most of them use expert system or Knowledge- based system (KBS) approach and some of them use Data mining approach. KBS benefits are more permanent, easier to duplicate, less expensive and automatically documented. Besides that, the limitations of KBS systems are difficult to capture informal knowledge; knowledge has not been documented and difficult to verbalize[9]. For those reason,

The most popular intelligent techniques for prediction are Artificial Neural Network, Decision Tree, Case-based Reasoning, Genetic Algorithm, Rough Set, Soft Computing (known as Hybrid Intelligent System), Operational Research and others techniques such as SVM, Fuzzy logic and etc. [28]. Basically, most of the prediction applications in Table

696

International Journal of Human and Social Sciences 5:11 2010

1 are used to predict stock, demand, rate, risk, event and etc., and a few apply on human or people. Besides that, the applications are mainly developed in business and industries fields, and only a limited discussion involved human resource in an organization.

Recently, with the advancement of technologies many researches use hybrid intelligent techniques to advance the capabilities of the single intelligent techniques. Most of them agree that hybrid technique can produce better results. In that case, they integrate the existing intelligent techniques, for example Fuzzy Artificial Neural Network, Fuzzy Data mining, Genetic Algorithm and Data mining and etc. In actual fact, in some cases that need the statistical concerns, the researchers integrate the intelligent techniques and statistical approaches to produce the best quality of results. For examples, the discrete choice theory and Data mining are used to predict the criminal events in the selected area; Data Mining and Fuzzy Artificial Neural Network is used to predict a suitable person for the project and etc. Table III shows some of the prediction applications that use hybrid intelligent techniques.

TABLE III

HYBRID TECHNIQUES FOR PREDICTION APPLICATIONS

Techniques used

Application

 

 

Soft Computing Adaptive Neuro-

Reservoir management [37]

Fuzzy Inference System (ANFSI) &

 

ANN)

 

Auto-regressive Moving Average

Service request [38]

(ARMA ) and Time delay neural

 

network (TDNN)

 

Neural networks and Case-based

Interest Rate [39]

reasoning

 

 

 

Discrete choice theory & Data

Criminal event[40]

mining

 

 

 

Data mining and mobile computing

Location in mobile

 

environment [41]

 

 

Genetic Algorithms (GA) and ANN

Financial forecasting [42]

 

 

ANN and Multivariate Adaptive

Breast cancer [43]

Regression Splines (MARS)

 

 

 

Data Mining and Fuzzy Artificial

Project Assignment[22]

Neural Network

 

 

 

III.KNOWLEDGE DISCOVERY IN DATABASE

A. Data Mining

Knowledge Discovery in Database (KDD) or Data mining (DM) is an approach that is now receiving great attention and is being recognized as a newly emerging analysis tool[28]. Data mining has given a great deal of concern and attention in the information industry and in society as a whole recently. This is due to the wide accessibility of enormous amounts of data and the important need for turning such data into useful information and knowledge[44]. Computer application such as DSS that interfaces with DM tool can help executives to make more informed and objectives decisions and help managers retrieve, summarize and analyze decision related data to make wiser and more informed decisions.

Data mining problems are generally categorized as clustering, association, classification and prediction[1, 10]. Over the years, the Data mining has involved various techniques including statistics, neural network, decision tree, genetic algorithm, and visualization techniques. Besides that, Data mining has been applied in many fields such as finance, marketing, manufacturing, health care, customer relationship and etc. Nevertheless, its application in HRM is rare[10].

B. Data Mining in HR Applications

Prediction applications in HRM are infrequent, as examples to predict the length of service, sales premiums, to persistence indices of insurance agents and analyze mis- operation behaviors of operators [10]. The research to date has listed researches in HRM problems domain uses DM approach. Table 4 lists some of the HR applications that use Data Mining, and it shows that there are few discussions about performance predictions that use DM technique in human resource domain.

TABLE IV

DATA MINING IN HR APPLICATIONS

Data Mining

Activity in HRM

method used

 

Fuzzy Data Mining

Employee development – Project

and Fuzzy Artificial

Assignment [22]

Neural Network

 

Decision tree

Personnel selection [10]

 

Job attitudes [20]

 

 

Association rule

Employee Development – Training

mining

[17]

 

 

Rough Set Theory

Personnel Selection – Recruit and

 

Retain Talents [21]

 

 

Fuzzy Data Mining

Personnel Selection [11]

 

 

IV. MANAGING TALENT IN HUMAN RESOURCE

MANAGEMENT

In an organization, talent management is becoming an increasingly crucial way of approaching HR functions. Talent is considered as any individual who has the capability to make a significant difference to the current and future performance of the organization[45]. In fact, managing talent involved human resource planning that regards processes for managing people in organization. Besides that, talent management can be defined as an outcome to ensure the right person is in the right job; process to ensure leadership continuity in key positions and encourage individual advancement; and decision to manage supply, demand and flow of talent through human capital engine[46]. The talent management process consists of recognizing the key talent areas in organization, identifying the people in the organization who constitute its key talent, and conducting development activities for the talent pool to retain and engage them and have them ready to move into more significant roles (see Fig.1).

There are several ways to identify talent in an organization and one of the methods is from employee performance records. The previous performance records

697

International Journal of Human and Social Sciences 5:11 2010

can be used to predict the right talent for the right job. The records can be analyzed using Data Mining method in order to find out the patterns and rules related to employee performance. The generated rules and patterns can perform a prediction model related to talent performance. These processes involve HR activities that need to be integrated into an effective system[47]. The decision support system can be use to forecast the right talent for the right job at the right time. In HRM, among the top current and future talent management challenges are developing existing talent; forecasting talent needs; attracting and retaining the right leadership talent; engaging talent; identifying existing talent; attracting and retaining the right leadership and key contributor; deploying existing talent; lack of leadership capability at senior levels and ensuring a diverse talent pool [3]. In this study, we focus one of the talent management challenges to identify the existing talent by predicting their performance.

employee performance, predict employee’s behavior and attitude, predicting performance progress, identify the best profile for different category of employee and etc. The matching of Data mining problems and talent management needs is very important, in a way to determine the suitable Data Mining techniques. In this study, we focus on identifying the patterns that relate to the talent by using prediction technique. This Data mining technique will produce prediction rules on how to identify talent for the right job.

Fig. 1 Talent Management Process

V. DATA MINING FOR TALENT MANAGEMENT

Data mining is among the best approach to analyze records in databases. The analyzed results can be use for future planning. From the literature that we discussed before, Data mining method also implemented in HR problem domains and most of researches in HR problems domain are focused on personnel selection task and few apply in other activities such as planning, training, managing talent and etc. Recently, the new demands and the increased visibility of HR management, HRM seeks a strategic role by revolving to Data Mining methods[1]. This can be done by identifying generated patterns from the existing data in HR databases as useful knowledge. The patterns can be generated by using some of the major Data mining techniques i.e. clustering, association, prediction and classification. There are many human resources tasks can be solved by using Data mining techniques such as employee evaluation, counseling techniques and performance management for effective and efficient decisions [1]. The tasks related to managing talent are summarized in Fig. 2.

For the example, by using prediction and classification techniques we can produce the percentage accuracy in

Fig. 2 Data Mining for Talent Management

In order to produce relevant Data Mining results that suitable to talent management tasks, there are several process in Data mining process should be followed. Fig. 3 gives us an overview of Data Mining task.

Talent data set selection

Cleaning and preprocessing for the related data

Identifying patterns

HR Database

Knowledge used

Patterns interpretation and

 

evaluation

HR Application

Fig. 3 Data Mining Process for Talent Management Tasks

According to Fig. 3, the first step is getting the main data set for Data mining. These may be collected from human resource operational databases or where the human resource data warehouse is selected. The selected data then goes through cleaning and preprocessing for removing

698

International Journal of Human and Social Sciences 5:11 2010

discrepancies and inconsistencies of data set and at the same time to improve quality of data set. Next, the data set is analyzed to identify patterns that represent relationship among data by applying algorithms, such as Neural nets, Decision Tree, Rough Set Theory and so on. Then patterns are validated with new human resource data sets. Besides that, it should be possible to transform the generated patterns into actionable plans that are likely to help the human resource people to achieve their goals. The steps in the mining process are repeated until meaningful knowledge is extracted. A pattern that satisfies these conditions becomes organizational knowledge and can be used in any related HR applications for talent management tasks.

A. Factors Related to Talent Forecasting

In this study, Data mining process will select data from the related human resources databases and transform it into useful knowledge. Data set selection process involve a study about the related competencies factors which can be identified from the actual data or databases. This is very important task in order to determine the significant attributes. Each employee has different competencies factors and that are depends on the type of their work. In this study we are focus on academician factors which is from Malaysian higher institutions. Fig. 4 shows some of the related factors for academic competencies. This involves process to determine the standard individual factors. Individual factors contains three main aspects; knowledge and expertise; management skill; and personal characteristics[48]. The academicians can also uses these individual factors which can be represented through academic context. Basically, as academician, they have to

do all the following tasks:

a)Professional qualification – this is a main activity for academicians, which are related to knowledge and expertise aspect. Each of academicians needs to focus their works on teaching, supervising, doing research and publication, organize and attend conferences and involve themselves in student activities. All these will contribute knowledge and expertise in their field.

b)Training – an academician should involve in any related training activities because these activities will upgrade their knowledge, management skill and coinciding can also contribute to their personal characteristics.

c)Administrative and contribution to university – usually in Malaysian higher institutions, most of the administration positions are occupied by the academicians. These positions involve a lot of decision making process, which are considered as a part of management skill and at the same time they will contribute to the achievement of university.

d)Social obligation – this activity usually necessitate something which are not related to their main job such as involve in any internal or external committees. The social obligation to the community will contribute to the development of university and country itself.

e)Award and appreciation – sometime an academician received honor for an achievement from their universities or community of interests towards their contributions. These are quite important in order to motivate them in their work and to develop the relationship between academicians with people coming from the outside.

Fig. 4 Academic Talent Factors

699

International Journal of Human and Social Sciences 5:11 2010

In order to achieve academician performance as an outcome of the performance evaluation process, academicians are demanded to prepare themselves with all this factors. Data selection is a beginning step in Data mining process, this step is a process to study attributes or parameters for each of the factors that could be represented as a group of data set.

B. Potential Data Mining Techniques

Data mining technique is the best linear unbiased estimator, decision tree and neural network is found useful in developing predictive models in many fields [28]. From the literature study, we list some of the characteristics for the popular intelligent techniques used in prediction application are listed in Table V.

TABLE V

POTENTIAL DATA MINING TECHNIQUES FOR PERFORMANCE PREDICTION IN

HR APPLICATION

Data Mining

Characteristics

Techniques

 

 

Artificial

Provide a variety of powerful tools for

optimization, function approximation, pattern

Neural

classification and modeling

Network(ANN)

 

 

 

 

1. Usually used for classification and

 

prediction tasks.

 

2. Produces a model which may represent

 

interpretable rules or logic statement and

 

more suitable for predicting categorical

 

outcomes

 

3. Non-parametric: suited to capturing an

Decision Tree

functional form relating independent and

dependent variables

 

4. Easy to interpret, computationally

 

inexpensive and capable to dealing with

 

noisy data. Model prediction is explainable

 

to the model user.

 

5. Automatic interaction detection – find

 

quickly significant high-order interactions

 

6. More informative outputs

 

 

 

Can explain and explore how the decision

Rough Set

was made with simple, understandable, and

useful rules in the presence or uncertainty

Theory

and vagueness

 

 

 

Fuzzy

More general than conventional methods

Used to construct relations among data and to

Clustering

transform relations into knowledge

 

 

 

Can be used for prediction as well as

Support Vector

classification and provide a compact

description of the learned model and highly

Machine (SVM)

accurate.

 

 

 

The techniques are also the potential techniques for prediction in HR applications. Some of the techniques are very well known among the researchers and they have proven as good prediction techniques. But in many cases it depends on the type of problems that they have. Table 5 lists some of the Data Mining prediction techniques that also suitable for the human resource data especially in managing talent such as Artificial Neural Network, Decision Tree, Rough Set Theory, Fuzzy Clustering and Support Vector Machine. In actual fact, Decision Tree has the advantages of easy interpretation and understanding for decision makers to compare with their domain knowledge for validation and

justify their decision [17]. Besides that, the Decision Tree can analyze various data without requiring the assumptions about the underlying distribution [10]. Recently, the hybrid methodology is another alternative to advance the capabilities of the application and to produce better results. The technique can also be hybrid with other techniques such as Artificial Neural Network (ANN), Knowledge-Based System, fuzzy logic, software agent and etc. In fact, some studies are needed to evaluate which technique is the best technique for human resource data, especially to handle issues in talent management. In that case, HR data that relate to talent such as age, position, race, gender, evaluation marks and etc. can be tested to several Data mining techniques in a way to find out the best technique. For that reasons, for future works we attempt to hybrid the suitable Data mining technique and knowledge-based system (KBS) approach for the system application. This integration can allow users to interact with the system and get the forecasting results and explaining about the decision made by the knowledge discovered from the database.

VI. HR SYSTEM ARCHITECTURE FOR TALENT FORECASTING

As a result from the literature study and the possibility to implement IDSS technology in human resource problem domains, we suggest the potential HR system architecture that uses past experience knowledge shown in Fig. 5. Basically, HR system architecture contains four main components:

a)Knowledge Discovery in Database (KDD) approach is used to develop predictive model and to find out the possible pattern and rules from existing database system. In this study, we will use HR databases that relate to talent performance such as personnel information, performance evaluation data and other related databases. The relevant data will be transformed into useful knowledge as predictive model, generated rules or classification of patterns. Decision tree and fuzzy logic techniques are among the potential intelligent techniques in this architecture.

b)Model Management System is a model base system that can store the constructed model, existing simulation model and related models that can be used in appropriate decision making process. In fact, before using talent performance predictive model, the model must be evaluated and tested in model analysis and evaluation process.

c)Knowledge Base System (KBS) contains a set of facts and rules. In the suggested architecture for HR application, KBS will contain information about talent patterns, association rules related to the potential talent in future and any related facts and rules. The rules and pattern will be evaluated and interpreted by the HR domain experts.

d)Advisory System is as inference engine in HR DSS application that supervises the interactions among the various parts of HR application. Basically, this component will react as interface between user and the

700

International Journal of Human and Social Sciences 5:11 2010

system itself, especially to display the prediction results, justify and explain the decision and sometimes if needed can instruct KBS to update the existing knowledge. In this study, the advisory system will display the potential talent with some reasons, and suggest the possible tasks for them.

This HR system architecture embedded KDD techniques with other DSS components such as knowledge-base and model-base. In this study, the suggested architecture is focusing on prediction purposes which are used to predict academician future performance. On the other hand, this architecture can also be applied to other Data Mining tasks such as association, classification and clustering in talent management tasks as mentioned in Fig. 2.

that, some of the techniques can be hybrid with other techniques which can produce better decision making results. Finally, the ability to continuously change and obtain new understanding is the power of HR application, and this can be the HR applications of future work.

VII. CONCLUSION

This article has described the significance of study, literature review on HR applications; prediction and intelligent techniques used; talent management concepts; related research in HRM known as HR application; intelligent techniques used in HR applications; the overview about KDD or Data Mining; and potential intelligent techniques for prediction. From the literature study, most researchers have discussed HR applications from different categories. However, there should be more HR applications that use intelligent techniques applied to different problem domains in HRM field research, in order to broaden our horizon of academic and practice work on HR applications. Due to these reasons, we propose HR system architecture for talent performance prediction based on past experience data; and it is developed towards problem-oriented domain. For future work, the data in HR can be tested using some of the prediction techniques to find out the suitable and best techniques, especially in HR problem domains. Besides

Fig. 5 Suggested HR System Architecture for Talent Forecasting

701

International Journal of Human and Social Sciences 5:11 2010

REFERENCES

[1]Ranjan, J., Data Mining Techniques for better decisions in Human Resource Management Systems. International Journal of Business Information Systems, 2008. 3(5): p. 464-481.

[2]DeNisi, A.S. and R.W. Griffin, Human Resource Management. 2005, New York: Houghton Mifflin Company.

[3]A TP Track Research Report Talent Management: A State of the Art. 2005, Tower Perrin HR Services.

[4]Stavrou-Costea, E., The challenges of human resource management towards organizational effectiveness A comparative study in Southern EU. Journal of European Industrial, 2005. 29(2): p. 112-134.

[5]DeCenZo, D.A. and S.P. Robbins, Fundamentals of Human Resource Management. 8th Ed. ed. 2005, New York: John Wiley & Son.Inc. .

[6]Okpara, J.O. and P. Wynn, Human resource management practices in a transition economy: Challenges and prospects. Management Research News, 2008. 31(1): p. 57-76.

[7]Hooper, R.S., et al., Use of an Expert System in a personnel selection process. Expert Systems and Applications, 1998. 14(4): p. 425-432.

[8]Hamidah, J., H. Abdul Razak, and A.O. Zulaiha. Potential Intelligent Techniques in Human Resource Decision Support System (HR DSS). in Proceedings 3rd International Symposium on Information Technology 2008. Kuala Lumpur: IEEE

[9]Martinsons, M.G., Knowledge-based systems leverage human resource management expertise. International Journal of Manpower, 1995. 16(2): p. 17-34.

[10]Chien, C.F. and L.F. Chen, Data mining to improve personnel selection and enhance human capital: A case study in high-technology industry. Expert Systems and Applications, 2008. 34(1): p. 380-290.

[11]Tai, W.S. and C.C. Hsu (2005) A Realistic Personnel Selection Tool Based on Fuzzy Data Mining Method. http://www.atlantis- press.com/php/download_papaer?id=46 9/1/2008.

[12]Huang, L.C., et al. Applying fuzzy neural network in human resource selection system. in Proceeding NAFIPS '04. IEEE Annual Meeting of the Fuzzy information 2004. 2004.

[13]Huang, L.C., et al., A neural network modelling on human resource talent selection. International Journal of Human Resource Development and Management, 2001. 1(Number 2-4): p. 206-219.

[14]Quintero, A., D. Konare, and S. Pierre, Prototyping an Intelligent Decision Support System for improving urban infrastructures management. European Journal of Operational Research, 2005. 162(3): p. 654-672.

[15]Qian, Z., G.H. Huang, and C.W. Chan, Development of an intelligent decision support system for air pollution control at coal-fired power plants. Expert System with Applications, 2004. 26(3): p. 335-356.

[16]Viademonte, S. and F. Burstein, From Knowledge Discovery to computational Intelligent : A Framework for Intelligent Decision Support System. 2006, London: Springer London.

[17]Chen, K.K., et al., Constructing a Web-based Employee Training Expert System with Data Mining Approach, in Paper in The 9th IEEE International Conference on E-Commerce Technology and The 4th IEEE International Conference on Enterprise Computing, E- Commerce and E-Services (CEC-EEE 2007). 2007.

[18]Mehrabad, M.S. and M.F. Brojeny, The development of an expert system for effective selection and appointment of the jobs applicants in human resource management. Computers & Industrial Engineering, 2007. 53(2): p. 306-312.

[19]Liao, S.-H., A knowledge-based architecture for implementing collaborative problem-solving methods in military e-training. Expert Systems and Applications, 2007. In Press(Corrected Proof).

[20]Tung, K.Y., et al., Mining the Generation Xer's job attitudes by artificial neural network and decision tree - empirical evidence in Taiwan. Expert Systems and Applications, 2005. 29(4): p. 783-794.

[21]Chien, C.F. and L.F. Chen, Using Rough Set Theory to Recruit and Retain High-Potential Talents for Semiconductor Manufacturing.

IEEE Transactions on Semiconductor Manufacturing, 2007. 20(4): p. 528-541.

[22]Huang, M.J., Y.L. Tsou, and S.C. Lee, Integrating fuzzy data mining and fuzzy artificial neural networks for discovering implicit knowledge. Knowledge-Based Systems, 2006. 19(6): p. 396-403.

[23]Glenzer, C., A conceptual model of an interorganizational intelligent meeting-scheduler (IIMS). Strategic Information Systems, 2003. 12(1): p. 47-70.

[24]Bozbura, F.T., A. Beskese, and C. Kahraman, Prioritization of human capital measurement indicators using fuzzy AHP. Expert Systems and Applications, 2007. 32(4): p. 1100-1112.

[25]Haddawy, P. and N.T.N. Hien (2007) A decision support system for

evaluating

international

student

applications.

http://www.apqn.org/event/past/details/102/presentation/files/6_prof_ peter_haddaway_and%20-hyuyen_thi_ngoc_hien.pdf 9/1/2008.

[26]Pardos, Z., et al. (2007) The effect of Model Granularity on Student Performance Prediction using Bayesian Networks. http://www.educationaldatamining.org/um2007/Pardos.pdf

[27]Sullivan, W.G. and W.W. Claycombe Technological Fundamentals of forecasting. 1977, Virginia: Reston Publishing Company, Inc.

[28]Tso, G.K.F. and K.K.W. Yau, Predicting electricity energy comsumption : A comparison of regression analysis, decision tree and nerural networks. Energy, 2007. 32: p. 1761 - 1768.

[29]Delen, D., G. Walker, and A. Kadam, Predicting breast cancer survivability: A comparison of three data mining methods. Artificial Intelligent in Medicine, 2005. 34(2): p. 113-127.

[30]Chang, L.Y. and W.C. Chen, Data mining of tree-based models to analyze freeway accident frequency. Journal of Safety Research 2005.

36(4): p. 365-375.

[31]Becerra-Fernandez, I., S.H. Zanakis, and S. Walczak, Knowledge discovery techniques for predicting country investmnet risk.

Computers & Industrial Engineering, 2002. 43(4): p. 787-800.

[32]Enke, D. and S. Thawornwong, The use of data mining and neural networks for forecasting stock market returns. Expert Systems with Applications, 2005. 29(4): p. 927-940.

[33]Hong, T. and I. Han, Knowledge-based data mining of news information on the internet using cognitive maps and neural networks.

Expert Systems with Applications, 2002. 23(1): p. 1-8.

[34]Liew, P.L., et al., Comparison of artificial neural networks with logistic regression in predicition of Gallbladder disease among obese patients. Digestive and Liver Disease, 2007. 39(4): p. 356-362.

[35]Lin, F.Y. and S. McClean, A data mining approach to the prediction of corporate failure. Knowledge-Based Systems, 2001. 14(3-4): p. 189-195.

[36]Cardoso, G. and F. Gomide, Newspaper demand prediction and replacement model based on fuzzy clustering and rules. An International Journal on Information Sciences, 2007. 177(21): p. 4799- 4809.

[37]Chang, F.J. and Y.T. Chang, Adaptive neuro-fuzzy inference system for prediction of water level in reservoir. Advances in Water Resources, 2006. 29(1): p. 1-10.

[38]Balaguer, E., et al., Predicting service request in support centers based on nonlinear dynamics, ARMA modeling and neural networks.

Expert Systems with Applications, 2008. 34(1): p. 665-672.

[39]Kim, S.H. and H.J. Noh, Predictability of Interest Rates using data mining tools : A comparative analysis of Korea and the US. Expert Systems with Applications, 1997. 13(2): p. 85-95.

[40]Xue, Y. and D.E. Brown, Spatial analysis with preference specification of latent decision makers for criminal event prediction.

Decision Support Systems, 2006. 41(3): p. 560-573.

[41]Yavas, G., et al., A data mining approach for location prediction in mobile environments. Data & Knowledge Engineering, 2005. 54(2): p. 121-146.

[42]Kim, K.-j., Artificial Neural Networks with evolutionary instance selection for financial forecasting. Expert Systems with Applications, 2006. 30(3): p. 519-526.

[43]Chou, S.M., et al., Mining the breast cancer pattern using artificial neural networks and multivariate adaptive regression splines. Expert Systems with Applications, 2004. 27(1): p. 133-142.

[44]Han, J. and M. Kamber, Data Mining : Concepts and Techniques. 2006, San Francisco: Morgan Kaufmann Publisher.

[45]Lynne, M., Talent Management Value Imperatives : Strategies for Execution. 2005, The Conference Board.

[46]Cubbingham, I., Talent Management : Making it real. Development and Learning in Organizations, 2007. 21(2): p. 4-6.

[47]CHINA UPDATE (2007) HR News for Your Organization : The Tower Perrin Asia Talent Management Study. http://www.towersperrin.com

[48]Chen, S.H. and H.T. Lee, Performance evaluation model for project managers using managerial practices. International Journal of Project Management, 2007. 25: p. 543-551.

702