Signal and Image Processing Letters

Vol. 4., No. 2, 2022, pp. 1-15

ISSN 2714-6677
https://simple.ascee.org/
1

Classification of interests and talents in early adult phase

based on RMIB test with neural network
Ahmad Azhari a,1,*, Erlangga Jaya a,2
a
Department of Informatics, Universitas Ahmad Dahlan, Yogyakarta, Indonesia
1
[email protected]; 2 [email protected]
* corresponding author

ARTICLE INFO ABSTRACT

The brain in the human body is responsible for regulating the overall work
of the human body and mind. The left part of the brain is the center of
intelligence or commonly called Intelligence Quotient (IQ). Intelligence
can come from genes received by children from their parents that will
continue to develop along with a person's maturity process. An individual
will go through a transition period or transition period, namely in the early
adult phase so that individuals in the early adult phase often experience
Keywords unstable psychic conditions. This labile condition occurs in early adult
Early Adult Phase individuals in this case, namely students who are still not sure what
EEG potential interests and talents they have, causing students It felt wrong to
Neural Network take the major. The purpose of this study is to classify interests and
Interests aptitudes from EEG data obtained from interviewees with RMIB test
Aptitudes stimulus. In this study, testing will be carried out on the object of study
RMIB Test where the object of study is an individual in the early adult phase with
an age range between 18-30 years. The test is carried out using a beta
signal (12-30 Hz) resulting from an Electroencephalogram (EEG) signal
filter generated from recording EEG data with the NeuroSky Mindwave
tool and then reduced to get the best value or component with the
Principal Component Analysis (PCA) method. EEG data recording is
carried out 3 times with data recording intervals every 14 days. EEG data
is information that we can get from activity waves in the brain, because
waves in the brain cannot be observed visually. Testing on this study.
The EEG data obtained will go through the pre-processing stage, namely
signal filters and signal reduction and then will be classified using neural
networks with a backpropagation algorithm with Using 1 layer of hidden
layer. In this study, the results of the RMIB test carried out by the
interviewees were calculated by psychologists (expert judgement) which
were used as comparison data or the output produced by the system.
Testing is carried out by cross validation, which is to cross-test each data
retrieval. Accuracy testing on the first fetch resulted in an accuracy of
92.8571%, in the second data retrieval it produced an accuracy of
78.571%, in the third data retrieval it resulted in an accuracy of 71.4285%
with an average accuracy produced by the system of 80.9523%.

This is an open access article under the CC–BY-SA license.

1. Introduction
The brain in humans is responsible for regulating the entire set of work of the human body and
mind. Therefore, that is the reason for the strong connection between the brain and the mind. The big
brain is divided into two parts, namely the right brain and the left brain. Experts say that the left
brain is the center of intellectual ability or Intelligence Quotient (IQ) [1], [2].

https://simple.ascee.org/ [email protected]
2 Signal and Image Processing Letters ISSN 2714-6677
Vol. 4, No. 2, 2022, pp. 1-15

The intelligence, interests and talents possessed by each individual are different. Intelligence can
be interpreted as a talent or ability possessed by an individual is innate obtained from his parents. The
level of intelligence can reflect the abilities of an individual and abilities can also come from other
people around that individual. Ability is directly proportional to interest, therefore if an individual
has an interest in a certain activity then the individual can like and pay attention to the activity with
a sense of pleasure and interest is an activity or activity that is fixed and is carried out to listen or pay
attention and keep memories of some of the activities carried out both intentional and unintentional
[3].
So that intelligence, interests and talents can be measured or known through electroencephalogram
(EEG) signals because EEG signals are bioelectrical signals that come from the surface of human
skin, EEG signals are complex and can be used as a source of information from what the brain does
or the functions of the brain. EEG signals are in the form of electrical waves that are very small in size
and different in each individual.
In this case, the intelligence test or commonly called the IQ test with the RMIB test question type
is a stimulus to obtain EEG data in the search for interests and talents from interviewees in the early
adult phase with an age range of 18-30 years. Where at this age is a transitional phase from adolescence
to adulthood (Natari, 2016). The reality that occurs today is not uncommon for individuals in the early
adult phase in this case, namely students who are still not sure what potential interests and talents they
have. So we also find various cases of students taking the wrong major because of the uncertainty that
arises in the chosen major with what has been lived is different from what they decided when choosing
the major to be pursued.
This problem is proven by the results of research by Youthmanual, Tech Incubator of Multimedia
Nusantara University that 92% of high school / K students and equivalent are confused and do not
know what major they will take in the future and 45% of students feel wrong in taking a major and
this research also found the main factor that causes the problem, namely they do not understand what
interests and talents or potentials they have. Meanwhile, the results of research by the Indonesia Career
Center Network (ICNN) stated that 87% of Indonesian students took the wrong major [4], [5].
In this study, EEG raw data obtained from interviewees who had done the RMIB test as a stimulus
would go through the pre-processing stage, namely the filter process into beta waves (13-30 Hz). The
next process is to reduce beta waves with the Principal Component Analysis method to get the best
value components and then in the process with a backpropagation artificial neural network to be
classified and then the results are compared with the RMIB test results that have been done by the
examiner and have been calculated by psychologists as experts.

2. Literature Review
2.1. Previous Studies
The research conducted by Ani and Hindarto (2017) resulted in the Extraction of EEG Signal
Characteristics for Epilepsy Disease Disorders Using the Wavelet Method. This study was conducted
because epilepsy occurs because there is a disturbance of the nervous system in the human brain,
which is recorded from the EEG signal [6].
2.2. Artificial Intelligence
Artificial intelligence is a part of computer science that studies how a computer or machine can be
made in such a way as to be able to do work like the ability of humans to get work done. Artificial
intelligence is a computer system formed to know and model human thought processes and design
machines to mimic human behavior [7].
2.3. Electroencephalogram (EEG)
EEG signals are bioelectrical signals that come from the surface of human skin, these EEG signals
are complex and can be used as a source of information from what the brain does or brain function.
EEG signals are in the form of electrical waves that are very small in size, so direct visual observation
is very difficult to do. EEG signals are recorded and processed to provide information on electrical
activity within the brain. The form of EEG signals in each person is different because it is influenced
by various types of influences such as emotional state, mental, age, activity and health conditions.
These EEG signals make it easier to provide information from the mind to an individual [8].

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)
ISSN 2714-6677 Signal and Image Processing Letters 3
Vol. 4, No. 2, 2022, pp. 1-15

The frequency of the dominant EEG signal wave generated will differ depending on the activities
carried out by one person to another. The first signal wave is Alpha (8-12Hz) with the activity carried
out is relaxation starting to rest or in the transition period from the conscious state to the unconscious
state. Beta waves (13-30Hz) are activities that are fully concentrated in, including in a state of
thinking, focus, and alertness. Theta waves(4-7Hz) individuals are in a mild sleep state, or very
drowsy. Delta waves (0.1-3Hz) are conditions in which an individual is fast asleep [9].
2.4. Neural Network
In this study using the neural network method, it is a paradigm of processing information inspired
by biological nerve cell systems, just like the human brain processing information. The element that
forms the basis of this paradigm is the new structure of an information processing system. Artificial
neural networks are just like humans, where they can learn from an example seen. Artificial neural
networks are formed to solve or provide solutions to specific problems such as pattern recognition
or classification resulting from the learning process [10]. As in Fig. 1. Fig. 1 is an artificial neural
network structure with a backpropagation algorithm.

Fig. 1. Neural network structure

2.5. Interests and Talents

According to the Big Indonesian Dictionary (KBBI) talent is the basis (intelligence, traits and
bearings) carried from birth. Talent is closer to the word aptitude which means carrying skills, which
is related to certain abilities and potentials possessed by an individual [11]. There is another opinion
that talent is a condition in an individual that allows him with a specific practice activity to achieve
abilities, knowledge and skills special [12].
Talents can be divided into 2 types, including:
1. General Talent
General talent is an ability that is a basic ability of a general nature. General talent can be defined
if all or many people have it.
2. Special Talents
Special talent is an ability that does mean special, not everyone has it.
2.6. Rothwell Miller Interest Blank (RMIB) Test
The Rothwell Miller Interest Blank Test (RMIB) is a type of inventory test in psychology. This
RMIB test focuses on assessing interests and aptitudes in an individual. The RMIB test is a
standardized test of potential interest and aptitude originally invented by Rothwell in 1947 and later
developed and updated by Miller in 1950.
In this RMIB test, there were initially 9 types of work on Rothwell's idea and then developed into
12 types of work by Miller and those types of work included, namely [13].
1. Outdoors 7. Literary
2. Mechanical 8. Musical
3. Computational 9. Social Service
4. Scientific 10. Clerical
5. Personal Contact 11. Practical
6. Aesthetic 12. Medical

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)
4 Signal and Image Processing Letters ISSN 2714-6677
Vol. 4, No. 2, 2022, pp. 1-15

3. Method
3.1. Object of Research
Because the purpose of this study is to find out the potential interests and talents in the early adult
phase, the object of this study is individuals with the age of 18-30 which in this case is students.
3.2. System Testing
In this study, system testing was carried out with several tests, namely:
1. Testing Training Data and Test Data
Testing of training data and test data is carried out with cross validation models, namely by creating
3 cross validation models.
2. Calculation of Accuracy Rate
Calculation of the accuracy rate using the confussion matrix, based on the values in Table 1 can
produce precission, recall and accuracy values

Table 1. Confussion matrix

Actual Value
TRUE FALSE
TP FP
TRUE (True Positive) (False Positive)
Predicted Correct result Unexpected result
Results FN TN
FALSE (False Negative) (True Negative)
Missing result Correct absence of result

4. Results and Discussion

After the text edit has been completed, the paper is ready for the template. Duplicate the template
file by using the Save As command, and use the naming convention prescribed by your conference
for the name of your paper. In this newly created file, highlight all of the contents and import your
prepared text file. You are now ready to style your paper; use the scroll down window on the left of
the MS Word Formatting toolbar.
4.1. EEG Data Collection
The initial stage is to collect EEG data from 14 speakers with an age range of 18-30 years which
will be used as training data and test data that have been obtained using the NeuroSky Mindwave tool
which have one electrode in the FP1 position to obtain the EEG data.
4.2. Identify the Headings
This stage is to carry out the process of filtering raw EEG data. This stage aims to separate the
beta signal (13-30 Hz) from other signals, because in this study the signal used is only a beta signal.
From Fig. 2 it can be explained that:
1. Filter the resulting signal
2. Type of response type, namely bandpass
3. Fill in the signal specification values to be made with the following details:
a. FS value = 512
b. Fstop1 value = 12.9
c. Fpass1 value = 13
d. Fpass2 value = 30
e. Fstop2 Value = 30.1

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)
ISSN 2714-6677 Signal and Image Processing Letters 5
Vol. 4, No. 2, 2022, pp. 1-15

Fig. 2. Bandpass filter

4.3. Signal Reduction

153600 beta signals that have been obtained from the raw data filter process will then be reduced
by the Principal Component Analysis (PCA) method to get the best value or component as input in
the system.

Table 2. Reduction yield weights

Data to - Data Weights Data to - Data Weights Data to - Data Weights
1. 14.00872 16. 1.89689 31. 0.35904
2. 6.011329 17. 1.05062 32. 0.35904
3. 4.999991 18. 2.510126 33. 0.35904
4. 3.994988 19. 0.059721 34. 0.35904
5. 3.036725 20. 0.07709 35. 0.35904
6. 2.960654 21. 0.244079 36. 0.359041
7. 2.042503 22. 0.376898 37. 0.35904
8. 2.02533 23. 0.35904 38. 0.359041
9. 1.921084 24. 0.358499 39. 0.359041
10. 5.9986 25. 0.849942 40. 0.35904
11. 8.050322 26. 0.35985 41. 0.35904
12. 3.1244 27. 0.384994 42. 0.991159
13. 1.34651 28. 0.358499
Total Weights
14. 6.26359 29. 0.35985
84.77%
15. 4.43951 30. 0.35904

In Table 2, it can be explained that the reduction results in the initial EEG data of 153600 to 42
data have a total weight of 84.77%, which means that the 42 data of the reduction results represent
84.77% of the initial 153600 data.
4.4. System Design
1. Flowchart Design
a. Fig. 3 flowchart classification of interests and talents

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)
6 Signal and Image Processing Letters ISSN 2714-6677
Vol. 4, No. 2, 2022, pp. 1-15

Fig. 3. Flowchart determination of potential interests and talents

b. Fig. 4 flowchart Method Neural Network Backpropagation

Fig. 4. Neural network backpropagation method flowchart

4.5. Implementation
1. Initial View
The initial view of the created system is as shown in Fig. 5 to Fig. 7.

Fig. 5. Initial view

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)
ISSN 2714-6677 Signal and Image Processing Letters 7
Vol. 4, No. 2, 2022, pp. 1-15

Fig. 6. Signal load process

Fig. 7. Filter results and signal reduction

No Source Code No Source Code

1. 2. function proses_Callback(~, ~, handles)
clear global filt;
global data;
d = fdesign.bandpass('N,Fst1,Fp1,Fp2,Fst2,C',3,11.9,20,30,30.1,512);
d.Stopband1Constrained = true;
d.Astop1 = 60;
function sinyal_Callback(~, ~, handles)
d.Stopband2Constrained = true;
clear global data;
d.Astop2 = 80;
[nama_file1, directory] =
Hd = design(d,'equiripple');
uigetfile({'*.txt'},'Open File');
filt = filter(Hd,data);
data = csvread([nama_file1 directory ]);
axes(handles.axes1);
axes(handles.axes);
save ('hasil_beta');
plot(data);
load ('hasil_beta','blanket');
global data;
filt2 = zscore(filt);
[COEFF SCORE] = princomp(filt2);
inputRMIB = SCORE(1:42,:);
plot (inputRMIB);
save ('hasil_reduksi');
global blanket;
Listing 1. Process load and signal filter

Listing Description 1:
a. The first line is for the data load process by retrieving files that have the format *.txt.
b. The second line is to filter the signal into a beta signal (13-30 Hz) and reduce it with PCA.

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)
8 Signal and Image Processing Letters ISSN 2714-6677
Vol. 4, No. 2, 2022, pp. 1-15

Fig. 8. Parameter Value Input and Test Data Selection

Fig. 9.Load Accuracy System

N Source Code N Source Code

o o
1. function proses_Callback(hObject, eventdata, handles) 1. inputRMIB(:,31),inputRMIB(:,32),inputRMIB(:,33),inputRM
g = str2double(get(handles.goals,'string')); IB(:,34)...
hl = str2double(get(handles.h_layer,'string')); inputRMIB(:,35),inputRMIB(:,36),inputRMIB(:,37),inputRM
ep = str2double(get(handles.epoch_,'string')); IB(:,38)...
load ('hasil_PCA.mat'); inputRMIB(:,39),inputRMIB(:,40),inputRMIB(:,41),inputRM
%%% value retrieval from radio button selection %%% IB(:,42)];
if get(handles.r1,'Value') xlRange = 'Q49:AD49';
target = else
[inputRMIB(:,1),inputRMIB(:,2),inputRMIB(:,3),inputRMIB target =
(:,4)... [inputRMIB(:,29),inputRMIB(:,30),inputRMIB(:,31),inputR
inputRMIB(:,5),inputRMIB(:,6),inputRMIB(:,7),inputRMIB(: MIB(:,32)...
,8)... inputRMIB(:,33),inputRMIB(:,34),inputRMIB(:,35),inputRM
inputRMIB(:,9),inputRMIB(:,10),inputRMIB(:,11),inputRMI IB(:,36)...
B(:,12.. inputRMIB(:,37),inputRMIB(:,38),inputRMIB(:,39),inputRM
inputRMIB(:,13),inputRMIB(:,14)]; IB(:,40)...
targetLatih = [targetData23]; inputRMIB(:,41),inputRMIB(:,42)];
dataLatih = targetLatih = [targetData12];
[inputRMIB(:,15),inputRMIB(:,16),inputRMIB(:,17),inputR dataLatih =
MIB(:,18)... [inputRMIB(:,1),inputRMIB(:,2),inputRMIB(:,3),inputRMIB
inputRMIB(:,19),inputRMIB(:,20),inputRMIB(:,21),inputRM (:,4)...
IB(:,22)... inputRMIB(:,5),inputRMIB(:,6),inputRMIB(:,7),inputRMIB(:
inputRMIB(:,23),inputRMIB(:,24),inputRMIB(:,25),inputRM ,8)...
IB(:,26)... inputRMIB(:,9),inputRMIB(:,10),inputRMIB(:,11),inputRMI

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)
ISSN 2714-6677 Signal and Image Processing Letters 9
Vol. 4, No. 2, 2022, pp. 1-15

inputRMIB(:,27),inputRMIB(:,28),inputRMIB(:,29),inputRM B(:,12..
IB(:,30)… inputRMIB(:,13),inputRMIB(:,14),inputRMIB(:,15),inputRM
inputRMIB(:,31),inputRMIB(:,32),inputRMIB(:,33),inputRM IB(:,16)...
IB(:,34)... inputRMIB(:,17),inputRMIB(:,18),inputRMIB(:,19),inputRM
inputRMIB(:,35),inputRMIB(:,36),inputRMIB(:,37),inputRM IB(:,20)...
IB(:,38)... inputRMIB(:,21),inputRMIB(:,22),inputRMIB(:,23),inputRM
inputRMIB(:,39),inputRMIB(:,40),inputRMIB(:,41),inputRM IB(:,24)...
IB(:,42)]; inputRMIB(:,25),inputRMIB(:,26),inputRMIB(:,27),inputRM
xlRange = 'C49:P49'; IB(:,28)];
elseif get(handles.r2,'Value') xlRange = 'AE49:AR49';
target = end
[inputRMIB(:,15),inputRMIB(:,16),inputRMIB(:,17),inputR % Prepare training data and training targets
MIB(:,18)... data_latih = [dataLatih];
inputRMIB(:,19),inputRMIB(:,20),inputRMIB(:,21),inputRM target_latih = [targetLatih];
IB(:,22)... [~,N] = size(data_latih);
inputRMIB(:,23),inputRMIB(:,24),inputRMIB(:,25),inputRM
IB(:,26)...
inputRMIB(:,27),inputRMIB(:,28)];
targetLatih = [targetData13];
dataLatih =
[inputRMIB(:,1),inputRMIB(:,2),inputRMIB(:,3),inputRMIB
(:,4)...
inputRMIB(:,5),inputRMIB(:,6),inputRMIB(:,7),inputRMIB(:
,8)...
inputRMIB(:,9),inputRMIB(:,10),inputRMIB(:,11),inputRMI
B(:,12..
inputRMIB(:,13),inputRMIB(:,14),inputRMIB(:,29),inputRM
IB(:,30)...

2. % JST process creation 3. %plot compare

net = newff(minmax(data_latih),[hl filename = 'Book1.ods';
1],{'logsig','purelin'},'traingdx'); sheet = 1;
% Provide value to influence the training process target_latih_asli = xlsread(filename, sheet, xlRange);
net.performFcn = 'mse'; figure,
net.trainParam.goal = g; plot(hasil_uji,'bX-')
net.trainParam.show = 20; hold on
net.trainParam.epochs = ep; plot(target_latih_original,'ro-')
net.trainParam.mc = 0.95; hold off
net.trainParam.lr = 0.1; grid on
% Training process title(strcat(['Output Graph NN vs Expert Judgement with
[net_keluaran,tr,Y,E] = train(net,data_latih,target_latih); value MSE = ',...
% Results after training num2str(error_MSE)]))
bobot_hidden = net_output. IW{1,1}; xlabel('Data to -')
bobot_keluaran = net_output. LW{2,1}; ylabel('Label to -')
bias_hidden = net_output.b{1,1}; legend('Output NN','Data Target','Location','Best')
bias_output = net_keluaran.b{2,1}; save ('hasil_uji1.mat')
jumlah_iterasi = tr.num_epochs;
nilai_keluaran = Y;
nilai_error = E;
error_MSE = (1/N)*sum(nilai_error.^2);
%%% test data %%%
data_test = [target];
hasil_uji = round(sim(net_output,data_uji))
4. function load_hasil_Callback(hObject, eventdata, handles)
load('hasil_uji1','hasil_uji')
set(handles.hasil_klasifikasi,'string',num2str(hasil_uji));
A = get(handles. expjud,'string');
B = get(handles.hasil_classification,'string');
xx = strsplit(A)
yy = strsplit(B)
b1 = yy(~cellfun(@isequal,yy,xx))
b2 = length(b1)
b3 = 14-b2
d = (b3/14)*100
set(handles.akurasi,'string',num2str(d));
Listing 2. Classification and testing process

Description:
a. The first row is a fork to specify the training data and the target data.
b. The second line is the process of creating artificial neural networks such as assigning
parameter values that affect the training process, training process and results after training.

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)
10 Signal and Image Processing Letters ISSN 2714-6677
Vol. 4, No. 2, 2022, pp. 1-15

c. The third line is the creation of a comparison graph plot between the system output and
the results of expert judgement.
d. The fourth line is a function to calculate the accuracy of the system.
4.6. Testing
Testing is carried out to conduct trials on systems that have been completed by applying them
using EEG data that has been obtained as many as 42 data from three data collections and the results
have been determined by experts to be used as a reference on the system can be seen in the Table 3.

Table 3. Cross validation

Cross Validation Test Results

P2 & P3 = Training data

CV 1 P1
P1 = Test Data
P1 & P3 = Training data
CV 2 P2
P2 = Test Data
P1 & P2 = Training data
CV 3 P3
P3 = Test Data

The explanation of the test stages as shown in Table 3 is as follows:

1. Cross Validation Testing
a. Cross Validation 1
1. The values of P1, P2 and P3 are stored in the input RMIB variables which are 42
x 42 in matrix size.
2. P2 and P3 as training data.
3. P1 as test data.
4. Determine expert result labels for P2 and P3 training data.
b. Cross Validation 2
1. The values of P1, P2 and P3 are stored in the input RMIB variables which are 42
x 42 in matrix size.
2. P1 and P3 as training data.
3. P2 as test data.
4. Determine expert result labels for P1 and P3 training data.
c. Cross Validation 3
1. The values of P1, P2 and P3 are stored in the input RMIB variables which are 42
x 42 in matrix size.
2. P1 and P2 as training data.
3. P3 as test data.
4. Determine expert result labels for P1 and P2 training data.
Test results can be seen in the Table 4.
In Table 5 The Number of iterations can be stated that neurons in the hidden layer worth 7 have
a better number of iterations compared to the number of iterations in the neuron value in the hidden
another layer.

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)
ISSN 2714-6677 Signal and Image Processing Letters 11
Vol. 4, No. 2, 2022, pp. 1-15

Table 4. Identification results

System Results
(Neuron Hidden Layer)
Beta Signal Expert
No
Reduction Judgement
6 7 8 9 10

1 1 1 1 1 1 1

2 3 3 3 3 3 3

3 4 4 4 4 4 4

4 6 6 6 6 6 6

5 2 2 2 2 2 2

6 7 2 2 2 2 1

7 2 2 2 2 2 2

8 9 9 9 9 9 9

9 6 6 6 6 6 6

10 2 2 2 2 2 2

11 4 4 4 4 4 4

12 9 9 9 9 9 9

13 2 2 2 2 2 2

14 2 2 2 2 2 2

15 1 1 1 1 1 1

16 2 2 2 2 2 2

17 5 5 5 5 5 5

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)
12 Signal and Image Processing Letters ISSN 2714-6677
Vol. 4, No. 2, 2022, pp. 1-15

System Results
(Neuron Hidden Layer)
Beta Signal Expert
No
Reduction Judgement
6 7 8 9 10

18 6 6 6 6 6 6

19 2 2 2 2 2 2

20 7 1 3 3 4 5

21 7 1 3 3 4 5

22 9 9 9 9 9 9

23 10 8 3 8 6 6

24 2 2 2 2 2 2

25 4 4 4 4 4 4

26 9 9 9 9 9 9

27 2 2 2 2 2 2

28 2 2 2 2 2 2

29 1 1 1 1 1 1

30 2 8 6 3 6 5

31 5 5 5 5 5 5

32 2 2 2 2 2 2

33 2 2 2 2 2 2

34 8 5 6 5 6 4

35 7 5 6 5 6 4

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)
ISSN 2714-6677 Signal and Image Processing Letters 13
Vol. 4, No. 2, 2022, pp. 1-15

System Results
(Neuron Hidden Layer)
Beta Signal Expert
No
Reduction Judgement
6 7 8 9 10

36 9 9 9 9 9 9

37 4 4 4 4 4 4

38 2 2 2 2 2 2

39 4 4 4 4 4 4

40 9 9 9 9 9 9

41 8 5 6 6 6 4

42 3 3 3 3 3 3

Table 5. Number of iterations

Number of iterations on a Hidden Layer (HL)
CV
HL = 6 HL = 7 HL = 8 HL = 9 HL = 10
CV 1 619 626 945 3219 5590
CV 2 3195 602 1203 573 423
CV 3 147 91 154 132 136

2. Calculation of Accuracy Rate

After conducting tests using the system then compared with the results of expert judgement and
comparing the results of the process that has been carried out, the accuracy of the test results with the
following calculations Table 6 and Table 7.

Table 6.Test result data

Test Data
Data Appropriate Non-Compliant
CV1 13 1
CV2 11 3
CV3 10 4

Table 7. Accuracy of test results

CV 1 CV 2 CV 3
13 + 0 11 + 0 10 + 0
× 100 % × 100% 100 %
1 13 + 0 + 0 + 1 11 + 0 + 0 + 3 10 + 0 + 0 + 4
= 92.857 % = 78.571 % = 71,4285 %
13 11 10
2 × 100 % = 100 % × 100% = 100% × 100 % = 100 %
13 + 0 11 + 0 10 + 0
13 11 10
3 × 100 % = 92,857 % × 100% = 78,571 % × 100 % = 71,4285 %
13 + 1 11 + 3 10 + 4

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)
14 Signal and Image Processing Letters ISSN 2714-6677
Vol. 4, No. 2, 2022, pp. 1-15

In Table 7, it is explained that in number 1, namely accuracy number 2 precision and number 3 is
recall.

5. Conclusion
The results of identifying potential interests and talents using the Neural Network Backpropagation
method resulted in the largest percentage of accuracy in the first retrieval data of 92.8571% using the
neuron value on the hidden layer = 7 which resulted in a better number of iterations compared to the
use of neuron values on the hidden layer others such as 6, 8, 9 and 10 while the average accuracy
produced by the system of the three antiquities on each take is 80.9523%. This result is directly
proportional to the problem of psychological tests regarding where the subject will remember the first
measurement and simply repeat the answer at the next measurement without thinking about it again.
The speaker will feel that the first measurement has already been made so that the next measurement
will answer not seriously. This causes the percentage of accuracy in each data retrieval to always
decrease. The application of the Neural Network Backpropagation method to the system created is
able to classify interests and talents from EEG data and is able to provide feature value information
on each data.
If you use the same stimulus continuously in multiple data retrievals, the data retrieval time interval
is extended again with a minimum of 15 days apart. Using other methods of EEG data extraction to
obtain values that are important for the testing and training process because of the very length of data
in EEG data will consume more time if directly through the training and testing process. Ensure that
the EEG data recorder is properly attached to the narrator during data recording so that the resulting
data is good. In taking EEG data, the interviewee must get a more conducive place for the work of his
stimuli such as the RMIB test or other tests because they are prone to getting a lot of noise on the data
obtained if the narrator is disturbed by what is around him.
References
[1] O. A. Sergeeva, R. De Luca, K. Mazur, A. N. Chepkova, H. L. Haas, and A. Bauer, “N-oleoyldopamine
modulates activity of midbrain dopaminergic neurons through multiple
mechanisms,” Neuropharmacology, vol. 119, pp. 111-122, 2017.
[2] R. M. Gourdine, “We treat everybody the same: race equity in child welfare,” Social work in public
health, vol. 34, no. 1, pp. 75-85, 2019.
[3] B. Mellers, et al., “The psychology of intelligence analysis: drivers of prediction accuracy in world
politics,” Journal of experimental psychology: applied, vol. 21, no. 1, p. 1, 2015.
[4] N. Putri, "Youthmanual: The Number of Students Who Choose the Wrong Major Is Still High." [Online].
Available: http://www.skystarventures.com/youthmanual-angka-siswa-yang-salah-pilih-jurusan-masih-
tinggi/. [Accessed: 01-Jan-2019].
[5] S. Agustina and B. Y. Cahyono, “Politeness and power relation in EFL classroom interactions: A study on
Indonesian learners and lecturers,” International Journal of Language and Linguistics, vol. 3, no. 2, pp. 92-
100, 2016.
[6] W. P. Ani, “Ekstraksi Ciri Sinyal EEG Untuk Gangguan Penyakit Epilepsi Menggunakan Metode
Wavelet,” MATICS: Jurnal Ilmu Komputer dan Teknologi Informasi (Journal of Computer Science and
Information Technology), vol. 9, no. 2, pp. 62-66, 2017.
[7] J. McCarthy and C. Gastmans, ”Moral distress: a review of the argument-based nursing ethics
literature,” Nursing ethics, vol. 22, no. 1, pp. 131-152, 2015.
[8] R. Karmila, E. C. Djamal, D. Nursantika Department of Informatics, and F. MIPA Universitas Jenderal
Achmad Yani Jl Terusan Jenderal Sudirman, "Identification of Concentration Level of EEG Signal With
Wavelet and Adaptive Backpropagation," Semin. Nas. Apx. Technol. Inf. August, no. August, pp. 1907–
5022, 2016.
[9] A. Siswoyo, Z. Arief, and I. A. Sulistijono, "Classification of Brain Signals Using Fuzzy Logic Methods
With Neurosky Mindset," pp. 119–128, 2014.

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)
ISSN 2714-6677 Signal and Image Processing Letters 15
Vol. 4, No. 2, 2022, pp. 1-15

[10] Yusran, "Implementation of Artificial Neural Networks (jst) to Predict Un Value Results Using
Backpropagation Method," JIT Terap. , vol. 9, no. 4, 2016.
[11] R. B. King, D. M. McInerney, and R. J. Pitliya, “Envisioning a culturally imaginative educational
psychology,” Educational Psychology Review, vol. 30, pp. 1031-1065, 2018.
[12] W. Meeus “Adolescent psychosocial development: A review of longitudinal models and
research,” Developmental Psychology, vol. 52, no. 12, 2016.
[13] S. Ampuni, N. Kautsari, M. Maharani, S. Kuswardani, and S. B. S. Buwono, “Academic dishonesty in
Indonesian college students: An investigation from a moral psychology perspective,” Journal of Academic
Ethics, vol. 18, no. 4, pp. 395-417, 2020.

Ahmad Azhari et.al (Classification of interests and talents in early adult phase based on RMIB test with neural network)