Paper id 29201413

International Journal of Research in Advent Technology, Vol.2, No.9, September 2014
E-ISSN: 2321-9637
27
An Efficient Classification Algorithms for Employee
Performance Prediction
V.Kalaivani1, Mr.M.Elamparithi2
M.Phil. Scholar, Department of Computer Science,
Assistant Professor, Department of PG Computer Applications, STC, Tamilnadu
Email: vanivelusamy@rediffmail.com1, Email: parithi_1980@yahoo.com2
Abstract-In any organization’s talent management is becoming an increasingly crucial method of approaching
HR functions. Talent management can be defined as an outcome to ensure the right person is in the right job.
Human talent prediction is the objective of this study. Due to that reason, classification and prediction in data
mining which is commonly used in many areas can also be implemented in this study. There are various
classification techniques in data mining such as Decision tree, Neural networks, Genetic algorithms, Support
vector machines, Rough set theory, Fuzzy set approach. This research has been made by applying decision tree
classification algorithms to the employee’s performance prediction. Decision tree is among the popular
classification technique which generates a tree and a set of rules, representing the model of different classes,
from a given data set. Some of the decision tree algorithms are ID3, C5.0, Bagging, Random Forest, Rotation
forest, CART and CHAID. In this study, C4.5, Bagging and Rotation Forest algorithms are used. Experiments
were conducted with the data collected from an institution which is implemented in WEKA tool.
Index Terms -Human talent, Data Mining, Classification, Decision tree, C4.5, Bagging, Rotation Forest.
1. INTRODUCTION
Human resource has become one of the main concerns
of managers in almost all types of businesses which
include private concerns, educational institutions and
governmental organizations [6]. Talent is considered
as any individual who has the capability to make a
significant difference to the current and future
performance of the organization. In fact, managing
talent involves 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. In
HRM, talent management is very important and need
some attentions from HR professional [1].
Data mining is a step in the KDD process
concerned with the extraction of patterns from the
data. Nowadays, there are some researchers on
solving HRM problems that uses Data mining
approach. Basically, most of the Data Mining
researches in HR problems domain focus on
personnel selection task and few apply in other
activities such as planning, training, talent
administration and etc. [1].
Recently, with the new demands and the
increased visibility of HR management, thus, HRM
seeks a strategic role by revolving to data Mining
methods. This can be done by identifying the patterns
that relate to the talent. The patterns can be generated
by using some of the major Data Mining techniques.
The matching of Data mining problems and talent
management needs are very important, in a way to
define the suitable Data Mining techniques.
2. DATA MINING AND ITS TECHNIQUES
Data mining is a collection of techniques for efficient
automated discovery of patterns in large databases.
That must be actionable so that they may be used in
an enterprise’s decision making process. Data mining
techniques provides a way to use various data mining
tasks such as classification, regression, time series
analysis, clustering, summarization, association rules
and sequence discovery, to find solutions for a
specified problem.
2.1 Classification
Classification involves finding rules that partition the
data into separate groups. The input for the
classification is the training data set, whose class
labels are previously known. Classification explores
the training data set and constructs a model based on
the class label, and intentions to allocate a class label
to the future unlabeled records. Since the class field is
well-known, this type of classification is known as
supervised learning. There are several classification
models such as Decision Tree, Genetic algorithms,
statistical models and so on.
2.2 Association rules
Association rule is the descriptive model of data
mining this enables us to establish association and
relationship between large and classified data items

E-ISSN: 2321-9637
28
based on certain attributes and characteristics. The
result of Association rules can help prevent failures by
some appropriate measures.
2.3 Clustering
Clustering is a method of grouping data into different
groups, so that the data in each cluster share similar
trends and patterns. Clustering creates a major class of
data mining algorithms. The algorithm tries to
automatically partition the data space into a set of
regions or clusters, to which the instances in the table
are assigned, either deterministically or probability-wise.
The objective of the process is to identify all
sets of similar instances in the data, in some best
manner.
2.4 Neural Networks
Neural Networks are a new paradigm in computing,
which involves developing mathematical structures
with the ability to learn. The methods are the result of
academic attempts to model the nervous system
learning. Neural networks have the significant
capability to derive meaning from complicated or
imprecise data and can be used to extract patterns and
detect trends that are too complex to be noticed by
either humans or other computer techniques.
2.5 Web Mining
Web mining is a specialized application of data
mining. Web mining is a technique to process data
available on Web and search for useful data. Web
mining enables us to determine web pages, text
documents, multimedia files, images and other kinds
of resources from web. Pattern extraction is a web
mining process to monitor the original or uploaded
web pages, extract information from them and
generate matches of a specific pattern with necessary
information specified by a user. The pattern extraction
process enables us to efficiently surf and access data
available on the web.
3. RELATED WORKS
Lipsa Sadath (2013) used Data Mining (DM)
techniques for automated intelligent decisions from
rich employee database for predictions of employee
performance implementing the finest KM strategies,
thus achieving stable HR system and brilliant
business.
Qasem et al. (2012) discussed data mining
techniques which were utilized to build a
classification model to predict the performance of
employees. They used Decision Tree for build the
classification model, where various classification
rules were generated. To validate the generated
model, more than a few experiments were conducted
using real data collected from several concerns.
Salleh et al. (2011) tested the influence of
motivation on job performance for state government
employees in Malaysia. The study showed a positive
relationship between affiliation motivation and job
performance. As people with higher affiliation
motivation and strong interpersonal relationships with
colleagues and managers tend to perform much better
in their jobs.
Jantan et al. (2010) present the study on how
the potential human talent can be predicted using a
decision tree classifier. By using this technique, the
pattern of talent performance can be identified
through the classification process. In that case, the
hidden and valuable knowledge discovered in the
related databases will summarized in the decision tree
structure. In this study, they use decision tree C4.5
classification algorithm to generate the classification
rules for human talent performance records. Finally,
the generated rules are evaluated using the unseen
data in order to estimate the accuracy of the
predication result.
Jantan et al. (2010) also propose the potential
data mining techniques for talent forecasting. Data
mining technique is the best balanced estimator,
decision tree and neural network and is found useful
in developing predictive models in many fields. In
this study, they attempts to use classifier algorithm
C4.5 and Random Forest for decision tree; and
Multilayer Perceptron (MLP) and Radial Basic
Function Network for neural network. They focus on
the accuracy of the techniques to find the suitable
classifier for HR data. The data are for management
and professional employees from higher education
institution.
Same authors Jantan et al. (2011) discussed
Human Resources (HR) system architecture to
forecast an applicant’s talent based on information
filled in the HR application and past experience, using
Data Mining techniques. The goal of the paper was to
find a way to talent prediction in Malaysian higher
institutions. So, they have specified certain factors to
be considered as attributes of their system, such as,
professional qualification, training and social
obligation. Then, several data mining techniques
(hybrid) where applied to find the prediction rules.
ANN, Decision Tree and Rough Set Theory are
examples of the selected techniques.
Chein and Chen (2006) have worked on the
improvement of employee selection, by building a
model, to predict the performance of newly
applicants. Depending on attributes selected from
their CVs, job applications and interviews. Their
performance could be predicted to be a base for
decision makers to take their decisions about either
employing these applicants or not. As a result for their
study, they found that employee performance is
highly affected by education degree, the school tire,
and the job experience.
Engel et al. (Elsevier-2014) exploits an
Approach to improve the performance of Weka, a
popular data mining tool, through parallelization on

E-ISSN: 2321-9637
29
GPU-accelerated machines. As a result, they observed
speedup levels of at least 49%; drastically decreasing
the time, the algorithm consumes to handle a dataset.
4. METHODOLOGY
This study has been made by applying decision tree
classification algorithms to the employee performance
prediction. A decision tree is a classification scheme
which generates a tree and a set of rules, representing
the model of different classes, from a given data set.
Experiments were conducted with the data collected
from an institution. C4.5, Bagging and Rotation
Forest algorithms are used in this system.
4.1 C4.5 Algorithm
The decision trees generated by C4.5 can be used for
classification, and for this reason, C4.5 is often
referred to as a statistical classifier. In building a
decision tree, we can compact with training sets that
have record with unknown attribute values by
estimating the gain, or the gain ratio, for an attribute
by in view of only those records where those attribute
values are available. The gain and gain ratio is
calculated as follows:
Gain(X, T) = Info (T) - Info(X, T) ---------Eq. (1)
Gain(X, T)
Gain Ratio(X, T) = ------------------ ----Eq. (2)
Split Info(X, T)
We can classify records that have unknown
attribute values by estimating the probability of the
various possible results. Unlike CART, which
generates a binary decision tree, C4.5 produces trees
with variable branches per node. When a discrete
variable is selected as the splitting attribute in C4.5,
there will be one branch for each value of the
attribute.
Algorithm
C4.5 builds decision trees from a set of
training data in the same way as ID3, using the
concept of information entropy. The training data is a
set of already classified samples.
Each sample consists of a p-dimensional vector
, where the represent
attributes or features of the sample, as well as the
class in which falls.
At each node of the tree, C4.5 chooses the
attribute of the data that most effectively splits its set
of samples into subsets enriched in one class or the
other. The splitting criterion is the normalized
information gain (difference in entropy). The attribute
with the highest normalized information gain is
chosen to make the decision. The C4.5 algorithm then
persists on the smaller sub lists.
4.2 Bagging Algorithm
Bagging decision trees, an early ensemble method,
builds multiple decision trees by repeatedly
resampling training data with replacement, and voting
the trees for a consensus prediction. However, there
are strong empirical indications that bagging and
random subspace methods are much more robust than
boosting in noisy settings [21].
Algorithm: Bagging
The bagging algorithm-create an ensemble of
models (classifiers or predictors) for a learning
scheme where each model gives an equality-weighted
prediction.
Input:
= D, a set of d training tuples;
= k, the number of models in the ensemble;
= a learning scheme (e.g., decision tree algorithm,
back propagation, etc.)
Output: A composite model, M*.
Method:
(1) For i=1 to k do//create k models:
(2) Create bootstrap sample,Di,by
sampling D with replacement;
(3) Use Di to derive a model,Mi;
(4) End for
To use the composite model on a tuple, X:
(1) If classification then
(2) Let each of the k models classify X and
return the majority vote;
(3) If prediction then
(4) Let each of the k models predict a value for
X and return the average predicted value;
4.3 Rotation Forest Algorithms
Rotation Forest is a recently proposed method for
building classifier ensembles using independently
trained decision trees. It was found to be more
accurate than bagging, AdaBoost and Random Forest
ensembles. Rotation Forest [19] draws upon the
Random Forest idea. The base classifiers are also
independently built decision trees, but in Rotation
Forest each tree is trained on the whole data set in a
rotated feature space. As the tree learning algorithm
builds the classification regions using hyper planes
parallel to the feature axes, a slight rotation of the
axes may lead to a very different tree. The effect of
rotating the axes is that classification regions of high
accuracy can be constructed with fewer trees than in
bagging and AdaBoost.
Rotation Forest Algorithm:
1. The feature set is randomly split into K subsets.
2. PCA is applied to each subset.
3. All principal components are retained.
4. Arrange the PCA coefficients in a matrix (rotation
matrix).

E-ISSN: 2321-9637
Cross-validation 10-folds
% accuracy
Algorithm used Training set % accuracy
C4.5 84.79%
30
5. Apply the rotation matrix to the data features.
6. Build each decision tree on the rotated training
data.
4.4 Datasets
Data is collected from the educational institutions. In
order to collect the required data, a questionnaire was
prepared and distributed to employees working in an
institution. The questionnaire was filled by 217
employees. After the questionnaires were collected,
the process of preparing the data was accomplished.
First, the information in the questionnaires has been
transferred to Excel sheets. Then, the types of data has
been reviewed and modified. These files are prepared
and converted to (ARFF) format to be compatible
with the WEKA data mining toolkit. The database
consists of 53 attributes and the input dataset is shown
in Fig 1.
Fig.1. Input Dataset in Text Editor
4.5 Experimental Results and Discussion
Decision tree algorithms were applied with WEKA
and the accuracy of the classification techniques with
cross-validation test is depicted in table 1.It is
observed that C4.5, Bagging algorithms have poor
accuracy compare with Rotation Forest algorithm and
not suitable for this problem domain due to the nature
of the data. This table shows an accuracy percentage
for C4.5, Bagging and Rotation forest algorithm with
cross-validation of 10 folds test option.
Table 1 Results of Decision Tree Algorithms with
cross-validation test
Algorithm used
C4.5 41.47%
Bagging 45.62%
Rotation forest 51.46%
The accuracy of the classification techniques with
Training set test is depicted in table 2.It is observed
that C4.5, Bagging algorithms have poor accuracy
compare with Rotation Forest algorithm and not
suitable for this problem domain due to the nature of
the data. This table shows an accuracy percentage for
C4.5, Bagging and Rotation forest algorithm with
Training set test option.
Table 2 Results of Decision Tree Algorithms with
Training set test
Bagging 75.57%
Rotation forest 100%
4.6 Analysis of Algorithm with Accuracy
The Fig 2 indicates that above specified three
algorithms has a different accuracy value. An
accuracy percentage for C4.5 algorithm with cross-validation
of 10 folds test option is 41.47% and the
training set test option is 84.79%.
An accuracy percentage for Bagging
algorithm with cross-validation of 10 folds test option
is 45.62% and the training set test option is 75.57%.
An accuracy percentage for Rotation forest
is 51.46% and the training set test option is 100%. In
above results, the Rotation forest algorithm has the
highest Accuracy value.
Fig.2. Analysis of algorithm with accuracy
4.7 Analysis of Algorithm with Time
The Fig 3 indicates that above specified three
algorithms has a different time. Time taken for C4.5

E-ISSN: 2321-9637
31
45.62
100
100
90
80
70
60
50
40
30
20
10
is 0.02 seconds and the training set test option is 0
seconds.
Time taken for Bagging algorithm with
cross-validation of 10 folds test option is 0.11 seconds
and the training set test option is 0.08 seconds.
Time taken for Rotation forest algorithm
with cross-validation of 10 folds test option is 1.87
seconds and the training set test option is 1.81
seconds. Consider the above four algorithm C4.5 has
the minimum time value for build the model.
Fig.3. Analysis of algorithms with
Time
5. CONCLUSION
Classification is the one of the hottest topics in the
area of data mining. The research activities on this
topic is reviewed hence, the survey guides the
researches to get an idea about the recent
advancements with Classification.
On working on performance, many attributes
have been tested, and some of them are found
effective on the performance prediction. For
companies managements and human resources
departments, this model can be used in predicting the
newly applicant personnel performance. Several
activities can be taken in this case to avoid any risk
related to hiring poorly performed employee.
In this paper, the parameter time value is
from 0 seconds to 1.87 seconds among three
algorithms. The accuracy value is from 41.47% to
100% for both the cross-validation and training set
test options among three algorithms.
From the performance analysis, each
algorithm has performed well. Consider the above
results Rotation forest algorithm has better
performance than other three algorithms because it
has maximum accuracy value 51.46% for cross-validation
and 100% for training set test option.
Therefore, the Rotation forest Algorithm is more
efficient algorithm for employee’s performance
prediction when compared to other two algorithms.
6. FUTURE WORK
This research has described the significance of the
study on the use of data mining classification
techniques for employee’s performance prediction.
The performance and efficiency of this research can
be improved through some future enhancements.
Some of the future enhancements that can be involved
in this research are:
· Other Decision Tree Classification algorithms
can be used to obtain better performance.
· Collect more proper data from several institutions
or companies.
REFERENCES
[1] Jantan et al. “Towards applying Data Mining
Techniques for Talent Management”, IPCSIT
vol.2, IACSIT Press, Singapore, 2011.
[2] Jantan, H., Hamdan, A.R. and Othman, Z.A.
(2010b). “Human Talent Prediction in HRM
using C4.5 Classification Algorithm”,
International Journal on Computer Science and
Engineering, 2(08-2010), pp. 2526-2534.
[3] Sivaram, Ramar 2010.“Applicability of Clustering
and Classification Algorithms for Recruitment
Data Mining,” International Journal of
Computer Applications, vol 4, pp.23-28, 2010
[4] Qasem A.Al-Radaideh, Eman AI Nagi “Using
Data mining techniques to build a Classification
model for predicting employees performance,”
IJACSA,vol 3,pp.144-151,2012.
[5] Al-Radaideh, Q. A., Al-Shawakfa, E.M., Al-
Najjar, M.I. (2006). “Mining Student Data
Using Decision Trees”, International Arab
Conference on Information Technology (ACIT
2006), Dec 2006, Jordan.
[6] Chein, C., Chen, L. (2006) "Data mining to
improve personnel selection and enhance human
capital: A case study in high technology
industry", Expert Systems with Applications, In
Press.
41.47
51.46
84.79
75.57
0
Accuracy %
Cross validation Training set
Test option
C4.5
Bagging
RF

E-ISSN: 2321-9637
32
[7] Kayha, E. (2007). “The Effects of Job
Characteristics and Working Conditions on Job
Performance”, International Journal of Industrial
Ergonomics, In Press.
[8] Salleh, F., Dzulkifli, Z., Abdullah, W.A. and
Yaakob, N. (2011). “The Effect of Motivation
on Job Performance of State Government
Employees in Malaysia”, International Journal
of Humanities and Social Science, 1(4), pp. 147-
154.
[9] Jantan, H., Hamdan, A.R. and Othman, Z.A.
(2010a) “Knowledge Discovery Techniques for
Talent Forecasting in Human Resource
Application”, International Journal of
Humanities and Social Science, 5(11), pp. 694-
702.
[10] Senol zafer erdogan, mehpare timor “A Data
Mining Application in a Student Database”.
Journal of aeronautics and space technologies,
vol 2, July 2005 pp.53-57
[11] Ruxandra PETRE “Data Mining for the Business
Environment,” Database Systems Journal, vol.
4, 2013.
[12] Lipsa Sadath “Data Mining:A Tool for
Knowledge Management in Human Resource,”
International Journal of Innovative Technology
and Exploring Engineering (IJITEE), Vol-2,
April 2013.
[13] Brijesh Kumar Bhardwaj, Saurabh Pal,” Data
Mining: A prediction for performance
improvement using classification,” International
Journal of Computer Science and Information
Security, Vol. 9, April 2011.
[14] Vaneet Kumar,Dr. Vinod Sharma,” Student’s
Examination Result Mining: APredictive
Approach,”International Journal of Scientific &
Engineering Research, Volume 3, November-
2012 .
[15] Sajadin Sembiring, M. Zarlis, Dedy Hartama,
Ramliana S, Elvi Wani, “Prediction Of Student
Academic Performance By An Application Of
Data Mining Techniques,” International
Conference on Management and Artificial
Intelligence IPEDR vol.6,2011, IACSIT Press,
Bali, Indonesia.
[16] Abeer Badr El Din Ahmed, Ibrahim Sayed
Elaraby, “Data Mining: A prediction for
Student's Performance Using Classification
Method,” World Journal of Computer
Application and Technology 2(2): 43-47, 2014
DOI: 10.13189/wjcat.2014.020203
[17] Yang Yanga, Suzanne S. Faridb, Nina F.
Thornhilla”Data mining for rapid prediction of
facility fit and debottlenecking
ofbiomanufacturing facilities”, Journal of
Biotechnology 179 (2014) 17–25(Elsevier)
[18] Tiago Augusto Engela, Andrea Schwertner
Charaoa,, Manuele Kirsch-Pinheirob, Luiz-
Angelo Steffenelc,“Performance improvement
of data mining in Weka through GPU
acceleration,”Procedia Computer Science 2014
,pp.93 – 100,Elsevier
[19] J. J. Rodr´ıguez, L. I. Kuncheva, and C. J.
Alonso. Rotation forest: A new classifier
ensemble method. IEEE Transactions on Pattern
Analysis and Machine Intelligence,
28(10):1619–1630, Oct 2006.
[20] G. K. F. Tso and K. K. W. Yau, "Predicting
electricity energy consumption: A comparison
of regression analysis, decision tree and neural
networks," Energy, vol. 32, pp. 1761-1768,
2007.
[21] J. Han and M. Kamber, Data Mining: Concepts
and Techniques. San Francisco: Morgan
Kaufmann Publisher, 2006.
[22] J. Ranjan, "Data Mining Techniques for better
decisions in Human Resource Management
Systems," International Journal of Business
Information Systems, vol. 3, pp. 464-481, 2008.
[23] 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.
[24] 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.
[25] 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.
[26] 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.
[27] 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.
[28] K.P.Soman, Shyam Diwakar, V.Vijay, Insight
into Data Mining theory and practice, Prentice-
Hall of India Private Limited, 2006.
[29] Arun K Pujari, Data Mining techniques,
Universities Press (India) Private Limited, 2001.
[30] G.K.Gupta, Introduction to Data Mining with
Case Studies, PHI Learning Private Limited,
2006.

Paper id 29201413

More Related Content

What's hot (19)

Viewers also liked (14)

Similar to Paper id 29201413 (20)

More from IJRAT (20)

Recently uploaded (20)

Paper id 29201413