JRSSEM 2022, Vol. 02, No. 4, 555 – 561

E-ISSN: 2807 - 6311, P-ISSN: 2807 - 6494

10.36418/jrssem.v2i04.313 https://jrssem.publikasiindonesia.id/index.php/jrssem

IMPLEMENTING NEURAL NETWORK ON DATA MINING

TO PREDICTING KEY PERFORMANCE INDEX FROM

EMPLOYEE

Albert Yakobus Chandra

1

Putri Taqwa Prasetyaningrum

2

Irfan Pratama

3

1,2,3

Department of Information System, Faculty of Information Technology Universitas Mercu

Buana Yogyakarta Indonesia

*

e-mail: albert.ch@mercubuana-yogya.ac.id

1

putri@mercubuanayogya.ac.id

2

irfanp@mercubuana-yogya.ac.id

3

*Correspondence: albert.ch@mercubuana-yogya.ac.id

1

Submitted

: 05 November 2022

Revised

: 16 November 2022

Accepted

: 25 November 2022

Abstract: Employee performance could help the company success to achieve its goals. In this era

predicting employee performance is a necessity for companies to gain success. In this research, we

presented a prediction of the employee performance index using a neural network. In this study,

we are using the Cross-industry standard process for data mining (CRISP-DM) as the base

framework for the data mining phase. The neural network classification method is employed to

create the prediction model. The result shows that using a neural network could get a confusion

matrix at 97.777 percent. This model was then applied to get insight from the performance index

that showed a couple of factors is has an important role in employee performance.

Keywords: data mining; key performance index; neural network.

Albert Yakobus Chandra

1

Putri Taqwa Prasetyaningrum

2

Irfan Pratama

3

| 556

INTRODUCTION

Information technology is a new field

that combines information science,

computer science, telecommunications,

and electronics. The evolution of

information technology involved several

steps leading to different developments

(Chitechi & Otanga, 2020). It is well known

that information technology affects

humans in almost every field, including

education, medicine, and especially in the

world of work or employment. Information

Technology is ubiquitous in the

industrialized world (March & Smith, 1995).

The use of information technology

demands that employees adapt quickly to

technology, which will obviously have

different impacts on employees. The

technology simplifies many functions,

which enhances performance and improves

professional satisfaction. The employee is

an organization most valuable asset. As a

result, employee performance has the

greatest impact on organizational

performance and ability to function. Some

employees perceive these as opportunities

and are more likely to use them as active

job aids to improve job performance and

satisfaction (Bala & Venkatesh, 2016). This

condition indicates that when an

organization decides to use information

technology to improve operational

performance. At this time, information

restores vacant and outdated capabilities

and contributes to the development of

performance skill of employee. Good

performance and employee efficiency are

one of the keys to successful organizations

(Rahmanidoust & Zheng, 2019). In order to

determine the level of employee

performance, we need to identify key

performance. Key Performance Indicators

(KPIs) are performance assessment tools

that determine the extent of achievement

of desired parameters in industrial

production lines, that is of great

significance to the success of the

manufacturing company. Key performance

indicators reflect departmental

performance (Singh, 2015). To define this

key performance this paper implementing

neural network on data mining.

Artificial neural networks or simply

"neural networks" as the connection model,

the distributed parallel processing model,

and the polymorphic neural system. Neural

networks reappeared in the mid-1980s

after major advances in neuroscience

(Mohaghegh et al., 1995). The simply

"neuron", and utilizes the massive

computer's parallel elements to achieve

high-performance speeds. Neural networks

are often used for statistical analysis and

data modeling, in which their role is

perceived as an alternative to standard

nonlinear regression or cluster analysis

techniques (B. Cheng and D. M.

Titterington, 1994). The terminology, they

all try to borrow the structure and how the

biological nervous system works based on

our current understanding of it (Herzog &

Almeida, 2018). Neural networks are widely

used for effective data mining,

transforming raw data into usable

information. Data mining has fueled the

research and development of methods and

algorithms for manipulating vast amounts

of data to solve real-world problems

(Aggarwal, 2011). Data mining deals with

the analysis of large and complex

557 | Implementing Neural Network On Data Mining To Predicting Key Performance Index

From Employee

databases, using machine learning and

statistical techniques to discover new,

useful and interesting knowledge. Data

mining is still in its infancy, but companies

in a wide range of industries, including

retail, finance, healthcare, manufacturing,

transportation, and aerospace, have

historically already used data mining tools

and techniques to use their data. increase.

Data mining uses pattern recognition

techniques and statistical and

mathematical techniques to sift through

stored information. Data mining analysts

can identify important facts, relationships,

trends, patterns, exceptions, and otherwise

overlooked helps identify possible

anomalies (Katiyaar & Sharma, 2012).

This paper is expected give solution for

the company leader to identify the

strengths and weaknesses in their

performance. Once the company leader

obtains Implementing Neural Network on

Data Mining to define Key Performance

Index, it will be easier to make decisions

that will have a great impact on the

organization or the company. Currently, the

neural network is very suitable to solve data

mining problems because of its good

features, adaptive self-organization,

parallel treatment, distributed storage, and

high fault tolerance. Key Performance

Indicators (KPIs) are generally used to

measure the performance of the

management process to recommend

appropriate future directions (de Andrade

& Sadaoui, 2017). The result using a neural

network, on data mining it will be more

efficient and the Key Performance Index

has a maximum level of accuracy. Neural

networks have high acceptance capacity for

noisy data and high precision and are

preferred in data mining (Krieger, 1996).

MATERIALS AND METHODS

This research following Cross-Industry

Standard Process for Data Mining (CRISP-

DM) methodology. CRISP-DM act as

framework that provide standardized steps

on how to develop a data mining project.

CRISP-DM has six phases which is: business

understanding, data understanding, data

preparation, modelling, evaluation and

deployment [21]. In this research we are

using CRISP-DM as framework and

customize it to match the process and

result we wanted on this research. The

research phases diagram based on

customize CRISP-DM as follow:

Figure 1. Research Phase Diagram

Data Understanding

The data from (

BRI Data Hackathon -

People Analytics | Kaggle

, n.d.) is used as

dataset in this research. There are total

2205 rows data in this dataset. The

attributes in this dataset shown as below:

Albert Yakobus Chandra

1

Putri Taqwa Prasetyaningrum

2

Irfan Pratama

3

| 558

Figure 2. Attribute from dataset

Data Preparation

In this research, the data is prepared to

be used on the process, start from handling

the missing values, remove the duplicate

rows and load the data to be processing on

the modeling using neural network.

1. Modeling

In this phase the neural network

algorithm is employed. The data contained

numerical data and continuous data. The

neural network multilayer perceptron was

used to train the dataset and making the

neural network classifier. Using python to

process the dataset with neural network

classifier we get accuracy for neural

network is 97.777 confusion matrix.

Figure 3. Neural network confusion matrix

RESULTS AND DISCUSSION

With implementing neural network

method on dataset, it shown classification

data based on indicators that affect the

performance of employees. The neural

network models need to be refined so as to

increase their performance. In choosing the

neural network, considerations should be

given to selecting a model based upon the

prediction risk. Figure 4 displays

performance the employee base on

gender, it shown best female at work 0.127

and best male at work 0.12 It's seems that

male are more productive than female. If

we see on the chart that has shown male

has got more productive than female.

Figure 4. performance based on Gender

Figure 4 shown significant difference

performance base on marital status,

already married at work 0,139 and have not

married at work 0.078. That significant

difference, married people are more

productive at workplace. Married people

have a family that depending on them, this

makes them more serious & productive at

workplace, this make sense.

559 | Implementing Neural Network On Data Mining To Predicting Key Performance Index

From Employee

Figure 5. Performance base on marital

status

Figure 3 shows performance base on

division, this dataset we have three division.

The employee on Mantri Kupedes have

percentage 50.4 % , Mantri KUR 46.1% and

the smallest percentage is 3.4% from

Mantri Briguna. Dataset the employee on

Mantri Kupedes have best performance

mor than Mantri KUR and Mantri Briguna.

Figure 6. Performance base on type of

division

Figure 6 shows surprisingly on

performance base on employee status, that

contract employee only have standard

performance compared to best permanent

employee 12.18%.

Figure 7. Performance base on Employee

status

Figure 7 shows performances base on

age with best age density above 0.12. Age

attribute showed a positive effect on

performance. This could be due to newly

working employees have energy and high

enthusiasm for work. On the other hand,

older employees may have had much

experience that would influence their

performance. It was observed that

employees between the age of 24 years

and 45 years showed better performance.

Figure 8. Performance base on age

Figure 8 shows best performance with

density above 1.6. Employees without

dependent prior to having one dependent

get the best performance from the dataset.

Because more and more dependents will

make employees much more serious in

pursuing their work targets to stay afloat.

Albert Yakobus Chandra

1

Putri Taqwa Prasetyaningrum

2

Irfan Pratama

3

| 560

Figure 9. Performance base on number of

dependences

Figure 9 shows Performance base on

GPA best employee have 0.175 – 0.200,

seems a reasonable predictor of effective

job performance – a high GPA signals the

individual has a considerable degree of

competence.

Figure 10. Performances base on GPA

Figure 10 shows The Training attribute

had a slight effect on performance. The

percentage below 12 months 11.9%, 12-24

months 14.2%, and above 24 22.4%. The

employee who above 24 months

professional training showed better

performance compared to those below 12

months and 12 - 24 months. Generally, the

number of employees who had

outstanding performance after the training

was higher than those below 12 months

and 12 - 24 months.

Figure 11. Performance base on training

CONCLUSIONS

In this research shown that from the

dataset with using neural network classifier

we cloud get insight from the data and

doing exploratory data analysis (EDA)

resulted from the neural network classifier.

Couple insight has important role on

employee performance index such as

gender, marital status, GPA, division and

training. Using neural network to process

the data shown that this technique is very

good at predicting key performance index

from employee.

REFERENCES

Aggarwal, C. C. (2011). An introduction to

social network data analytics. In

Social

network data analytics

(pp. 1–15).

Springer.

B. Cheng and D. M. Titterington. (1994).

“RNA2. pdf,.”

Statistical Science

,

9

(1),

2–54.

Bala, H., & Venkatesh, V. (2016). Adaptation

to information technology: A holistic

nomological network from

implementation to job outcomes.

Management Science

,

62

(1), 156–179.

561 | Implementing Neural Network On Data Mining To Predicting Key Performance Index

From Employee

BRI Data Hackathon - People Analytics |

Kaggle

. (n.d.). Retrieved October 30,

2022, from

https://www.kaggle.com/competition

s/bri-data-hackathon-pa

Chitechi, K. V., & Otanga, D. A. (2020).

Current Trends In Information

Technology: Which Way For Modern It

Experts

.

de Andrade, P. R. M., & Sadaoui, S. (2017).

Improving business decision making

based on KPI management system.

2017 IEEE International Conference on

Systems, Man, and Cybernetics (SMC)

,

1280–1285.

Herzog, H., & Almeida, P. (2018).

Introduction to PET. In

Hybrid MR-PET

Imaging

(pp. 129–146).

Katiyaar, V., & Sharma, I. K. (2012). Use of

Data Mining & Neural Network in

Commercial Application.

Vinodhini

Katiyaar et al,/(IJCSIT) International

Journal of Computer Science and

Information Technologies

,

3

(3), 4041–

4049.

Krieger, C. (1996).

Neural Networks in Data

Mining

.

March, S. T., & Smith, G. F. (1995). Design

and natural science research on

information technology.

Decision

Support Systems

,

15

(4), 251–266.

Mohaghegh, S., Arefi, R., Bilgesu, I., Ameri,

S., & Rose, D. (1995). Design and

development of an artificial neural

network for estimation of formation

permeability.

SPE Computer

Applications

,

7

(06), 151–154.

Rahmanidoust, M., & Zheng, J. (2019).

Evaluation of Factors Affecting

Employees’ Performance Using

Artificial Neural Networks Algorithm:

The Case Study of Fajr Jam.

International Business Research

,

12

(10), 86–97.

Singh, K. (2015). Leadership Style and

Employee Productivity: A Case Study

of Indian Banking Organizations.

Journal of Knowledge Globalization

,

8

(2).

for possible open access publication

under the terms and conditions of the Creative

Commons Attribution (CC BY SA) license

(https://creativecommons.org/licenses/by-sa/4.0/).