Designing a Smart Prediction Model for Influence of the

Infrastructure Completeness on Work Satisfaction

Sukhbir Singh, Ermi Girsang, Sri Lestari R. Nasution

Faculty of Medical, Universitas Prima Indonesia, Indonesia

Keywords: Completeness of Infrastructure Facilities, Job Satisfaction, Employees.

Abstract: Based on Indonesian data, 47.1% of hospital permanent employees were found to be dissatisfied at work.

Dissatisfaction can be caused by several factors such as internal, external or a combination of both. The

external factors that highly suspected to cause work dissatisfaction were incomplete infrastructure facilities.

The purpose of this study was to design a prediction model of the influence of the infrastructure completeness

on work satisfaction from a population of 230 respondents with 70 samples (by stratified random sampling).

Statistically, questionnaire data were analyzed using univariate, bivariate with chi-square tests, and

multivariate with multiple logistic regression at a 95% confidence level ( = 0.05). Based on statistical test

values, the prediction model was built using the Adaptive Neuro-Fuzzy Inference System (ANFIS) method.

Of the six variables tested, found three variables (i.e., workspace, furniture / furniture, office equipment) that

significantly influence the works satisfaction, p <0.05. The workspace variable was the most dominant which

has an opportunity of 10.494 times higher to be satisfied with a poor workspace. Accuracy results of 98.7%

towards the design of predictive models was achieved.

1 INTRODUCTION

The hospital is a labor-intensive organization with

diverse employee resource backgrounds (Supriyanto

& Ernawati, 2015). Human resources at the hospital

are divided into 3 (three) groups, namely

professional, managerial and workforce. The

professional group is tasked with trying to cure

treated patients (Soeroso, 2016). All categories of

human resources in the hospital will have different

job satisfaction.

Job satisfaction is a form of employee perception

that is reflected in attitudes and focused on behavior

towards work. According to Kreitner & Kinicki

(2016), job satisfaction is a positive feeling about

one's work that is the result of an evaluation of its

characteristics.

Job satisfaction as a very important factor of

productivity and quality of work, especially in health

workers (Dragana, Arandjelovic, Maja, & Stanković,

2018). Research conducted by Jaiswal et al. (2015) in

India found that the average hospital employee job

satisfaction index was in the same range, but was

found to be highest for nurses (68%), followed by

doctors (66%), support staff (63%) and technicians (

62%). Research Dragana et al. (2018) in Sweden that

most employees think that their work is interesting

and stimulating, so they work enthusiastically. More

than 50% of health workers surveyed stated that they

were not satisfied at work. Research at the Jakarta

Hajj Hospital by Sulistyarini (2018) found that 47.1%

of the hospital's permanent employees were

dissatisfied with work. Prayoga, Lailiyah, & Sari's

(2017) research at the Blambangan District General

Hospital in Banyuwangi Regency states that all

hospital staff have a level of satisfaction in the

satisfied category.

Many ways you can do to measure someone's job

satisfaction. Researchers at Cornell University

developed the Smith, Kendall, & Hulin (1969)

approach called Job Descriptive Index (JDI) to assess

job satisfaction with several work dimensions,

namely work, salary, promotion, infrastructure

facilities, supervision, and colleagues. In terms of job

satisfaction, Gilmer (1966) in As'ad (2018) mentions

the factors that influence job satisfaction are

opportunities for advancement, job security, salary,

company and management, intrinsic and work

factors, working conditions, social aspects of work ,

communication and facilities.

The completeness of work facilities and facilities

will encourage the emergence of effective and

112

Singh, S., Girsang, E. and R. Nasution, S.

Designing a Smart Prediction Model for Inﬂuence of the Infrastructure Completeness on Work Satisfaction.

DOI: 10.5220/0010290401120117

In Proceedings of the International Conference on Health Informatics, Medical, Biological Engineering, and Pharmaceutical (HIMBEP 2020), pages 112-117

ISBN: 978-989-758-500-5

efficient work results and encourage quality

improvement in line with existing work standards.

Work facilities provided by the company must be

adjusted to the needs of the organization, so that the

work done by employees can run effectively

(Hasibuan, 2016; Turnip et al, 2020; Wijaya et al,

2019).

In this study, a preliminary survey was conducted

by interviewing 20 Hospital employees about

satisfaction with the completeness of facilities and

infrastructure in supporting work, as many as 13

people said that they were satisfied with the

infrastructure used and 7 people said they were not

satisfied. Dissatisfaction is caused because according

to respondents there are some infrastructure facilities

that are incomplete when they work so that it

interferes with the work implementation. Equipment

that is felt by the employee is lacking such as

ventilators, special beds, and others. The incomplete

infrastructure is also caused by old age such as

photocopiers, and others. Infrastructure such as

narrow parking, medical committee rooms and a less

ergonomic Central Sterile Supply Department (CSSD

/ sterilization) cause employees to be less satisfied.

Design model to predict the influence of the

infrastructure completeness on work satisfaction

from a population of 230 respondents with 70

samples (by stratified random sampling) is

performed.

2 METHOD

This type of research is a quantitative analytic study

with a cross sectional study design. The study was

conducted at the Stella Maris Hospital in Medan in

December 2019. The study population was all

permanent employees in the Hospital about 230

people, and samples were obtained as many as 70

respondents. Univariate data analysis, bivariate using

chi-square test, and multivariate using multiple

logistic regression tests with a confidence level of

95% ( = 0.05).

Adaptive Neuro Fuzzy Interference System

(ANFIS) is an algorithm that combines fuzzy systems

with artificial neural networks (Wijaya & Suhartono,

2012; Turnip et al, 2018). The basis of the integration

is the advantages and disadvantages of each system.

ANFIS was first introduced by Jang in 1993 (Jang,

1993). Neural networks recognize patterns and adapt

patterns to environmental changes, while fuzzy logic

combines human knowledge and seeks conclusions to

make a decision. The main advantage of artificial

neural networks is that they can recognize the system

through a learning process to improve adaptive

parameters. The weakness of this system is the

complexity of the structure. While the fuzzy system

has a concept similar to the concept of human

thinking. The combination of the two will

complement each other's strengths and weaknesses

(Kusumandari et al, 2018; Turnip et al, 2018).

Fuzzy inference system used is a first order

Tagaki-Sugeno-Kang (TSK) fuzzy inference system

with consideration of simplicity and computational

ease. The basis for rules with two fuzzy if-then rules

as below.

𝐼𝑓 𝑥



𝑖𝑠 𝐴



𝑎𝑛𝑑 𝑥



𝑖𝑠 𝐵



𝑇ℎ𝑒𝑛 𝑦



 𝑐



𝑥



 𝑐



𝑥



𝑐



𝑓 𝑥



𝑖𝑠 𝐴



𝑎𝑛𝑑 𝑥



𝑖𝑠 𝐵



𝑇ℎ𝑒𝑛 𝑦



 𝑐



𝑥



 𝑐



𝑥



𝑐



where x

and x

are inputs and A

, A

, B

are

degrees of membership with predetermined fuzzy

rules. Whereas c

+ c

and c

+ c

are linear parameters. ANFIS architecture can

be seen in the Figure 1. From the Figure1, it can be

seen that there are 5 layers or 5 layers in ANFIS

architecture. The neurons in the first layer are called

adaptive to the parameters of an activation function.

The output is in the form of a new degree of

membership that is formed from existing inputs,

namely 𝜇 𝐴



,𝜇 𝐴



,𝜇 𝐵



,𝜇 𝐵



. The membership

function used in this study was trimf.

Figure 1: ANFIS diagram with 5 layers.

The neurons in the second layer are fixed neurons

and use the And operator. The output is the product

of the degree of membership in layer 1 as

𝑤



𝜇 𝐴



.𝜇 𝐵



where 𝑤



is an α predicate or fixed neuron. Each

node output states the firing strength of each fuzzy

rule. This function can be expanded if the premise

section has more than two fuzzy sets. The number of

vertices in this layer shows how many rules are

formed. The neuron in the third layer is formed from

the calculation of the ratio of α predicates or fixed

Designing a Smart Prediction Model for Inﬂuence of the Infrastructure Completeness on Work Satisfaction

113

neurons from the i-th rule to the sum of all α

predicates. This output is often referred to as

normalized firing strength.

𝑤







𝑤



𝑤



 𝑤



 …  𝑤



If more than two rules are formed, the function

can be expanded by dividing w

by the total number

of w for all rules. On the third layer, the resulting

output becomes adaptive neurons. There are new

parameters called consequent parameters 𝑐



𝑥





𝑐



𝑥



𝑐



. This parameter is affected by the α

predicate. In this last layer there is only one output

node which is the output or the result of all the

calculations that have been done above.

𝑤





.𝑓





∑

𝑤



.𝑓



∑

𝑤



where

∑

𝑤





.𝑓



is y as an output system.

3 RESULTS AND DISCUSSIONS

Characteristics of respondents which most of them

were aged <32 years by 54.3%, a small proportion

aged> 32 years 45.7%. Based on gender, the majority

of respondents were 74.3% female. Based on

education, the majority of respondents had a diploma

education of 55.7%, the remaining are undergraduate

and master degree about 1.4%. Based on the length of

work, the majority of respondents worked> 5 years

about 64.3% and <5 years about 35.7%.

Based on the results of bivariate analysis, all

independent variables were significantly related to

employee job satisfaction (p = 0.016), work space (p

= 0.009), lighting (p = 0.043), furniture / furniture (p

= 0.001), communication tools (p = 0.027 ), office

supplies (p = 0.005), and air fresheners (p = 0.030).

Full Chi-Square statistical test results can be seen in

Table 1.

The results of multivariate analysis with multiple

logistic regression tests showed that as many as 7

variables as a candidate model, obtained as many as

3 variables that affect job satisfaction of hospital

employees, namely workspace, furniture / office

furniture and office equipment.

The most influential variable in this study is the

workspace variable which has a value of Exp (B) / OR

= 10.494 meaning that employees who declare a good

hospital workspace, have the opportunity to feel

satisfied by 10.4 times higher than employees who

declare workspace less well.

Table 1: Relationship of Each Independent and Dependent

Variable.

Table 3: Multiple Logistic Regression Test Results.

Variables B Sig. Exp(B)

95%CI

for

Exp(B)

Workspace

Furniture

Office

supplies

Constant

2,351

2,160

2,076

-9,641

0,006

0,003

0,009

0,000

10,494

8,671

7,970

1,965-

56,051

2,103-

35,757

1,665-

38,148

3.1 Workspace Effects

The results showed that there was an influence of

workspace on job satisfaction of hospital employees.

Employees who stated that the hospital workspace

was good, had the opportunity to feel satisfied by

10.494 times higher than employees who stated the

workspace was not good.

Job satisfaction is one factor that is able to

improve the performance of an employee so it needs

attention. Low conditions of job satisfaction can

cause employees to get bored with their tasks so that

sooner or later it is not reliable. Equipment or

Variables

Job Satisfaction

p-value

Satisfied

Less

satisfied

Building:

Good

Less

0,016

Workspace:

Good

Less

0,009

Lighting :

Good

Less

0,043

Furniture:

Good

Less

0,001

Communication

tool:

Good

Less

0,027

Office supplies:

Good

Less

0,005

Air Freshener:

Good

Less

0,030

HIMBEP 2020 - International Conference on Health Informatics, Medical, Biological Engineering, and Pharmaceutical

114

infrastructure is very supportive in the work to

facilitate and expedite work, especially work space.

The results of this study prove that the

completeness of workspace infrastructure facilities

affects employee job satisfaction. Employees who

state that the work space is complete and in good

condition tend to feel satisfied at work. The

management realizes that employee participation

must be considered because employees as human

resources play the most important and potential role

for hospital success. Completeness of infrastructure

related to work space in the form of area, cleanliness,

comfort and layout of photos or paintings.

With the completeness of the furniture provided,

the employee will feel satisfied at work. Biasana

satisfaction refers to the pleasure and love of his

work. In addition, employees also demonstrate

discipline by complying with predetermined rules and

demonstrate work performance both individually and

in groups.

3.2 Furniture Effects

It was found that there is influence of furniture on

employee job satisfaction in hospitals. Staff who

stated that the hospital furniture was good, had a 8.6

times higher chance of being satisfied compared to

those who did not.

The results of this study prove that furniture has a

significant effect on employee job satisfaction.

Employees who state that furniture is complete and in

good condition tend to feel satisfied compared to

incomplete ones. The completeness and availability

of furniture in the employee's office is related to the

arrangement of the location of furniture which is well

organized and neat, ergonomic chairs and tables,

cupboards to store adequate files, furniture conditions

that have long been replaced. All of that is to support

employee work and increase job satisfaction.

3.3 Office Supplies Effects

Based on the results of the study showed that there is

an influence of office equipment on job satisfaction

of hospital employees. Employees who state good

hospital office equipment have a 7.9 percent higher

chance of being satisfied compared to less good ones.

It was found that the completeness of the office

affects employee job satisfaction. Employees who

state that complete office equipment tends to feel

satisfied at work. The work environment can support

work improvement, the office was more attentive and

is able to provide office equipment to support the

convenience of employees at work. The availability

of office equipment needed by employees will

improve and accelerate work to work more

effectively and efficiently.

3.4 Smart Model with ANFIS

Data processing starts from entering data, designing

input and output forms, trying out training data, and

continuing with testing the training data. Next, the

model adjusted the range of values and shapes of the

input pyramid. 40% of the total data is used as

training data and the rest is used as test data. Figure 2

(a) shows the selected training data. Inputs, outputs,

and rules on anfis are set using Grid Partition.

Member Functions (MF) serves as a reference for the

distance value from the input that will categorize the

level of assessment of the respondent. In this

experiment, 3 MF to provide 3 categories of

assessment based on the value of the respondents

were used. For the output itself, a constant value is

used to increase the accuracy of the data to be

processed. In the end some rules that are related to

each other will be formed by themselves.

The accuracy of the modeling results is improved

by using a hybrid method. The use of hyrid is a

combination of backpropagation and least-squares

regression which aims to adjust the FIS parameters.

Error tolerance is given a value of 0 which indicates

that training will stop when the amount of training is

reached. The value of the epochs is given 10 which

indicates the data will be given 10 repetitions in order

to get maximum accuracy. Figure 2 (b) shows the

results of testing toward the training data. The results

on the data (blue dot) show that the test and training

data are coincidental (red dot) which indicate the high

level of modeling accuracy. It shows the data that we

present with our target of fulfilling what we want.

Figure 3 is a form of rule structure that has been

composed of various numbers in the data. The data

consists of 7 inputs in the form of Building, work

space, lighting, furniture, communication tools,

office supplies, and air fresheners and 1 output in the

form of employee job satisfaction. The rules formed

are interconnected with all input and possibilities

formed to produce output.

As explained in the previous Grid Partition, the

model will be formed as shown in Figure 3. Consists

of three pyramids that indicate the level of

respondents' assessment to be processed. Consists of

less, medium, and good, each of which represents the

category for the number obtained from the

respondent's satisfaction value. The distance value

used is 1 - 5. The use of this trimf model is very

suitable for this data because the accuracy obtained is

quite good.

Designing a Smart Prediction Model for Inﬂuence of the Infrastructure Completeness on Work Satisfaction

115

The process in Figure 3 is useful for determining

the level of satisfaction of employees based on the

results or the total number of questions from

respondents. Consists of 7 entries, each based on a

total of scores in one rating category. After the data is

tried and matched with the initial training data, then

the data is compared between the data from the

questionnaire and the data from the data that we have

designed. This is to prove whether anfis accuracy

system is qualified to be used. Figure 4 is the

displayed of developed rules. Finally, the accuracy of

97,6% of smart predictive model was obtained.

Figure 2: Training vs Testing data.

Figure 3: The design smart model for job satisfaction.

Figure 4: Rules viewer for parameter adjusting.

4 CONCLUSIONS

Workspace, furniture, office equipment affect

employee job satisfaction while building variables,

lighting, communication tools, air fresheners do not

affect job satisfaction. The most influential variable

in this study is the workspace variable which has an

Exp (B) / OR value of 10.494, which means that

employees who declare a good hospital workspace,

have the opportunity to feel satisfied by 10.494 times

higher than the less good. Smart prediction model for

influence of the infrastructure completeness on work

satisfaction with 97.6% accuracy is achieved. The

high level of accuracy indicates that the obtained

model can be used by management to predict the job

satifaction without repetition measurement data.

REFERENCES

Adisaputro, G., 2015. Manajemen Pemasaran Analisis

Untuk Perancangan Strategi Pemasaran (Cetakan 2).

Yogyakarta: UPP STIM YKPN.

As’ad, M., 2013. Psikologi Industri, Seri Ilmu Sumber

Daya Manusia (Cetakan 2). Jakarta: Liberty.

Dahlius, A., & Ibrahim, M., 2016. Pengaruh Fasilitas

Kerja Terhadap Kepuasan Kerja Karyawan Pada PT.

Bank Riaukepri Cabang Teluk Kuantan Kabupaten

Kuantan Singingi. JOM FISIP, 3(2), 11–40.

Dragana, N., Arandjelovic, M., Maja, N., & Stanković, A.

2008. Job satisfaction in health care workers. Acta

Medica Medianae, 47(1), 45–51.

Gilmer, V. H., 1966. Industrial Psychology (1st editio).

USA: McGraw Hill Book CompanyIn.

Hasibuan, M., 2014. Manajemen Sumber Daya Manusia

(Cetakan 4). Jakarta: Bumi Aksara.

Jaiswal, P., Gadpayle, A., Modi, R., Padaria, R., Singhal,

A., Sachdeva, S., & Vajala, R., 2015. Job satisfaction

among hospital staff working in a Government teaching

hospital of India. Medical Journal of Dr. D.Y. Patil

University, 8(1), 131.

Kreitner, R., & Kinicki, A., 2014. Perilaku Organisasi

(Edisi 9). Jakarta: Salemba Empat.

Kusumandari, D., Risqyawan, M., Yazir, M., Turnip, M.,

Darma, A. and Turnip, A., 2018. Application of

convolutional neural network classifier for wireless

arrhythmia detection, Journal of Physics: Conference

Series, Volume 1080 (2018) 012048 doi:

10.1088/1742-6596/1080/1/012048.

Moekijat, 2014. Manajemen Personalia dan Sumber Daya

Manusia (Cetakan 7). Jakarta: Pustaka Utama.

Pangarso, A., Firdaus, F. F., & Moeliono, K. N., 2016.

Pengaruh Fasilitas Kerja Terhadap Kepuaasan Kerja

Karyawan Divisi Sumber Daya Manusia dan Diklat PT.

Dirgantara Indonesia. Jurnal Administrasi Bisnis,

12(1), 50–62.

HIMBEP 2020 - International Conference on Health Informatics, Medical, Biological Engineering, and Pharmaceutical

116

Prayoga, D., Lailiyah, S., & Sari, J. D. E., 2017. Analisis

Kepuasan Karyawan Dan Akreditasi Rumah Sakit

Umum Daerah Blambangan Kabupaten Banyuwangi.

Jurnal Riset Akuntansi Dan Bisnis Airlangga, 2(2),

269–289.

Putro, B. A., 2013. Hubungan Antara Kinerja, Sarana,

Prasarana Dengan Kepuasan Kerja Pegawai Di

Kantor Dinas Pendidikan Kabupaten Blitar. Fakultas

Ilmu Pendidikan UM.

Smith, P. C., Kendall, L., & Hulin, C. L., 1969. The

Measurement of Satisfaction in Work and Retirement.

Chicago: Rand McNally.

Soeroso, S., 2016. Manajemen Sumber Daya Manusia di

Rumah Sakit. Suatu Pendekatan Sistem (Cetakan 2).

Jakarta: EGC.

Sulistyarini, N. (2013). Faktor-Faktor yang Berhubungan

dengan Kepuasan Kerja Pegawai Tetap di Rumah Sakit

Haji Jakarta Tahun 2013 (Fakultas Kedokteran dan

Ilmu Kesehatan UIN Syarif Hidayatullah Jakarta).

Supriyanto, S., & Ernawati. (2015). Pemasaran Industri

Jasa Kesehatan (Cetakan 2). Yogyakarta: Andi Offset.

Thomas, Y. A., Rorong, A. J., & Tampongangoy, D.

(2016). Pengaruh Fasilitas Kerja Terhadap Kinerja

Pegawai Negeri Sipil Di Kantor Dinas Pendidikan

Minahasa Tenggara. Universitas Sam Ratulangi

Manado.

Turnip, A., Andrian, Turnip, M., Dharma, A., Paninsari,

D., Nababan, T., Ginting, C.N., 2020. An application

of modified filter algorithm fetal electrocardiogram

signals with various subjects, International Journal of

Artificial Intelligence, vol. 18, no., 2020.

Turnip, A., Ilham Rizqywan, M., Kusumandari, D., et al.,

2018. Classification of ECG signal with Support

Vector Machine Method for Arrhythmia Detection,

Journal of Physics: Conference Series, Vol. 970

(2018) 012012 doi: 10.1088/1742-6596/970/1/012012.

Turnip, A., Kusumandari, D., Pamungkas, D., 2018. Drug

Abuse Identification based EEG-P300 Amplitude and

Latency with Fuzzy Logic Calssifier, IEEE International

Conference on Applied Engineering, (ICAE), 3-4 Oct.

2018, DOI: 10.1109/INCAE.2018.8579378.

Wijaya, C., Andrian, M., Harahap, M., Turnip, A., 2019.

Abnormalities State Detection from P-Wave, QRS

Complex, and T-Wave in Noisy ECG, Journal of

Physics: Conference Series, Volume 1230, (2019)

012015. doi:10.1088/1742-6596/1230/1/012015.

Designing a Smart Prediction Model for Inﬂuence of the Infrastructure Completeness on Work Satisfaction

117