Assessing the effectiveness of data mining tools in classifying and predicting road traffic congestion

Abstract

Traffic congestion is a significant issue in cities, impacting the environment, commuters, and the economy. Predicting congestion is crucial for efficient network operation, but high-quality data and computational techniques are challenging for scientists and engineers. The revolution of data mining and machine learning has enabled the development of effective prediction methods. Machine learning (ML) approaches have shown potential in predicting traffic congestion, with classification being a key area of study. Open-source software tools WEKA and Orange are used to predict and classify traffic congestion. However, there is no single best strategy for every situation. This study compared the effectiveness of both data mining tools for predicting congestion in one of the areas of the capital of the Hashemite Kingdom of Jordan, Amman, by testing several classifiers including support vector machine (SVM), K-nearest neighbors (KNN), logistic regression (LR), and random forest (RF) classifications. The results showed that the Orange mining tool was superior in predicting traffic congestion, with a prediction accuracy of 100% for Random forest, logistic regression, and 99.8% for KNN. On the other hand, results were better in WEKA for the SVM classifier with an accuracy of 99.7%.

Full text

Indonesian Journal of Electrical Engineering and Computer Science
Vol. 34, No. 2, May 2024, pp. 1295~1303
ISSN: 2502-4752, DOI: 10.11591/ijeecs.v34.i2.pp1295-1303  1295
Journal homepage: http://ijeecs.iaescore.com
Assessing the effectiveness of data mining tools in classifying
and predicting road traffic congestion
Areen Arabiat, Muneera Altayeb
Department of Communications and Computer Engineering, Faculty of Engineering, Al-Ahliyya Amman University,
Amman, Jordan
Article Info
ABSTRACT
Article history:
Received Jan 29, 2024
Revised Feb 9, 2024
Accepted Feb 16, 2024
Traffic congestion is a significant issue in cities, impacting the environment,
commuters, and the economy. Predicting congestion is crucial for efficient
network operation, but high-quality data and computational techniques are
challenging for scientists and engineers. The revolution of data mining and
machine learning has enabled the development of effective prediction
methods. Machine learning (ML) approaches have shown potential in
predicting traffic congestion, with classification being a key area of study.
Open-source software tools WEKA and Orange are used to predict and
classify traffic congestion. However, there is no single best strategy for
every situation. This study compared the effectiveness of both data mining
tools for predicting congestion in one of the areas of the capital of the
Hashemite Kingdom of Jordan, Amman, by testing several classifiers
including support vector machine (SVM), K-nearest neighbors (KNN),
logistic regression (LR), and random forest (RF) classifications. The results
showed that the Orange mining tool was superior in predicting traffic
congestion, with a prediction accuracy of 100% for Random forest, logistic
regression, and 99.8% for KNN. On the other hand, results were better in
WEKA for the SVM classifier with an accuracy of 99.7%.
Keywords:
K-nearest neighbor support
Logistic regression
Machine learning
Orange data mining tool
Random forest
Vector machine
WEKA data mining tool
This is an open access article under the CC BY-SA license.
Corresponding Author:
Areen Arabiat
Department of Communications and Computer Engineering, Faculty of Engineering
Al-Ahliyya Amman University
Al-Saro, Al-Salt, Amman, Jordan
Email: a.arabiat@ammanu.edu.jo
1. INTRODUCTION
Traffic congestion is one of the most important problems that residents of capital cities around the
world suffer from. It can lead to increased stress, delayed delivery, fuel waste, and financial losses. From this
standpoint, studies that contribute to reducing this traffic phenomenon are extremely important [1]. Most
modern studies of congestion forecasting are based on analyzing peak traffic periods, where forecasting is
classified into three types according to traffic flow: short-term, medium-term, and long-term forecasting.
Short-term forecasts which last between five and fifteen minutes on average have a lot of random volatility,
high complexity, and poor data stability. Given the complexity of the traffic condition, it is imperative to
provide reliable short-term forecasting for real-time information determination, On the other hand, medium-
and long-term forecast units often extend to days, weeks, months, and years, and because of the huge time
lag, the stability of the data is very high; Therefore, this type of forecasting is often used to estimate long-
term traffic flow with high accuracy through time series that rely on past data and expected future data [2].
Recent advances in traffic congestion prediction have given rise to an important topic of study, particularly in
AI and ML. The vast availability of data aided by navigation systems and fixed sensors has contributed to a

 ISSN: 2502-4752
Indonesian J Elec Eng & Comp Sci, Vol. 34, No. 2, May 2024: 1295-1303
1296
significant expansion of this subject of study over the past several decades as traffic data can be analyzed,
pattern recognized, and traffic flow insights using ML techniques [3], [4]. Accurate forecasting contributes to
traffic flow volume control, traffic management, and optimization, as predicting traffic congestion using ML
is considered more accurate than traditional methods, which contributes significantly to improving traffic
flow, especially at peak times. However, to fully utilize the potential of machine learning in traffic
management, issues such as data reliability and model interpretability must be resolved [5], [6].
The researchers in [7], used monitoring-based data from IoT sensors embedded in smart cities to
develop traffic control systems that operate autonomously and reduce traffic jams, to obtain great accuracy
and low error rates, a neuro-fuzzy algorithm was used. In validation testing, the model outperformed previous
techniques with an accuracy rate of 98%. To create a traffic prediction model, this study uses radio
frequencies. The RF algorithm features great accessibility, high stability, and outstanding reliability.
Liu and Wu in [8] presented an automated system to predict expected traffic using an RF classifier. The RF
algorithm features great accessibility, high stability, and outstanding reliability. The traffic forecast proposed
model is created using input data including weather, time of day, season, unusual road conditions, traffic
conditions, and holidays. The results showed that the traffic prediction model created using the RF
classification approach can be predicted effectively, has a modest generalization error, and has an accuracy
rate of 87.5%. Li et al. [9] presented a model using an RF classification technique to predict traffic
congestion state in another study. The random forest method has a reputation for being practical, flexible, and
highly effective. Weather, time of day, unique road conditions, road quality, and holidays were used as model
input factors to build road traffic forecasting models. As a result, the results show that the traffic prediction
model developed using the random forest classification method can be predicted successfully with an
accuracy of 87.5% with low generalization error. It is also more adept at predicting crowded scenarios due to
its fast calculation speed.
On the other hand, researchers in [10] developed a long short-term memory (LSTM) network as a
means of predicting congestion propagation across road networks, where the model predicts congestion
propagation over 5 minutes in Buxton, UK, which is a congested city. The study used both univariate and
multivariate LSTM models, with the former relying entirely on the speed recorded over the past five minutes
while the latter takes traffic flow rate and vehicle progress into account. The accuracy of the models ranged
between 84-95% depending on the route configuration and these results revealed that both models may
produce adequate prediction of congestion propagation over short periods, with accuracy mostly determined
by the topology of the local road network. The researchers in [11] developed a proposed model for accurate
prediction of short-term traffic conditions in smart transportation. Intelligent transportation systems (ITS)
systems using machine learning classifiers including LD-SVM, decision forests, MLP, and CN2 rule
induction. According to the results, decision forests outperformed other methods with an average
improvement of 0.982 and 0.975, respectively. This method solves the problem of overfitting in existing
modelling methodologies. Ratra and Gulia [12], presented a comparison of different techniques using
empirical and parameter analysis to evaluate two open data mining tools, WEKA and Orange. The results
reveal that WEKA outperforms Orange in terms of the qualities required for a fully functional and easy-to-
use rating platform. WEKA is suitable for data mining classification challenges. Additionally, the study
analysed proactivity and recall across datasets, and found that Orange had 82.4% greater proactivity and
80.6% greater recall. WEKA has greater precision (83.7%) and recall (83.7%). This comparison includes
Naïve Bayes, Random Forest, and nearest neighbor classifiers. The precision value of the k-nearest algorithm
is larger, with WEKA having a precision of 75.3% and a recall of 75.2%. This work presents a unique
comparison between two distinct data mining tools, where a large data set containing approximately 8,671
records was tested and the accuracy of the evaluations was approximately 100%. This unique experiment for
this study aims to determine the volume of traffic flow through one of the most traffic-congested districts in
the Jordanian capital, Amman.
2. METHOD
By testing the efficiency of several classifiers, the model proposed in this work provides
a comparison between the WEKA and Orange data mining tools, to measure the effectiveness of both tools
for predicting traffic congestion in the Jordanian capital, Amman, as the Greater Amman Municipality
provided the traffic data that was used in the classification process. A system architecture for predicting
traffic Congestion is shown in Figure 1, where the basic steps are stated. The first and most important stage
in developing a machine learning classifier is pre-processing to improve the quality of the dataset, making it
ready to feed a machine learning model. In order to predict traffic congestion in this work, the classifiers RF,
SVM, KNN, and LR were employed to find the optimum data mining technique and classifier for traffic

Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 
Assessing the effectiveness of data mining tools in classifying and predicting … (Areen Arabiat)
1297
congestion prediction. The results will be evaluated using confusion matrices such as accuracy, sensitivity,
precision, and F-measure.
Figure 1. System architecture for predicting traffic congestion
2.1. Dataset
The Greater Amman Municipality provided the data set for King Abdullah Street in Amman, in
2018. This data included time, date, traffic flow, capacity, number of lanes, road width, and traffic volume.
This data was collected with high accuracy using detectors and sensors that can count the number of passing
vehicles on a lane and calculate the traffic volume for each lane approach every hour of every day, every
month, for the whole year [13].
2.2. Data preprocessing
Data preprocessing is the process of removing or correcting erroneous, incomplete, or incorrect data
from a dataset. Use excel's remove duplicates tool to get rid of unnecessary data, then use conditional
formatting to fix any structural issues. To prepare the traffic dataset for use in building an ML classifier using
WEKA and Orange data mining tools, it is first saved as a CSV file.
2.3. Classification
Data classification is done in two steps: i) training, sometimes called learning and ii) testing, or
evaluation, when an instance's predicted class is compared to its actual class. If the hit rate is considered
acceptable by the analyst, then the classifier is considered capable of classifying future occurrences of
unknown classes. Normal and congested are the two categories into which the data in this investigation
should be divided. The model will be validated using 10-fold cross-validation with 30% of the data used for
testing and 70% of the data used for training through RF, SVM, KNN, and LR classifiers.
2.3.1. Random forest
High-dimensional datasets cannot be effectively used with the RF-supervised learning algorithm. An
infinite forest is created by combining M numbers of different decision trees [14]. Random forests use
decision trees arranged randomly on information units, creating forecasts and determining the best solution

 ISSN: 2502-4752
Indonesian J Elec Eng & Comp Sci, Vol. 34, No. 2, May 2024: 1295-1303
1298
through voting. This method provides a useful insight into trait significance [15]. A composite classifier
generates multiple decision trees and integrates them for efficient outcomes. The random forest model uses
decision trees trained on random characteristics but typically ignores the diverse contributions of trees in
different test instances. The aggregate that each tree provides individually is averaged to make predictions.
Also, incredibly adept at adjusting to sacrifice is random forest [16], [17]. In classification issues, the towing
rule, deviation, and Gini index are the primary rules used to binary divide data of these guidelines, the most
often applied is the Gini index in (1), which quantifies the node impurity:
  
󰇛  
󰇜

 (1)
The target class is A, and the sample fraction of class an is 
. A node with a modest value of μ is
considered to be pure, meaning that it has good class separation and mostly comprises observations from a
single class [18]. Figure 2 depicts the RF structure, Where X = X1, X2, X3,..., XN, and n is the number of data
dimensions or predictive variables, which is an example of an input data set, While T expresses trees T1(X),
T2(X), T3(X), Tn(X) that form the RF model [18].
Figure 2. Random forest (RF) structure [19]
2.3.2. Logistic regression
Gaussian-form numerical input variables are used in the binary classification process when an
outcome in regression modeling is binary or dichotomous (yes/no). This is a specific case that is known as
logistic regression, where each input value has a coefficient that is then transformed via a logistic function.
This fast method works well for a variety of classification problems [20]. Assuming a straight-line
relationship between independent variables, linear regression is a frequently used kind of regression analysis
for continuous outcomes. It is useful for determining how one independent variable affects a continuous
result. Nevertheless, it is preferable to use multivariate linear regression to find distinct contributions while
concurrently assessing the effects of several components [21]. The logistic regression model has a particular
form that is described in (2):
󰇛󰇜
󰇛󰇜 
  ( β0+ β1 X1+……+ βk Xki) (2)
where (1-Pi) is the chance that Y takes a value of 0, Pi is the probability that Y takes a value of 1, and e is the
exponential constant [22].
2.3.3. Support vector machine
The machine learning technique known as SVM is grounded in statistical learning theory, as its
algorithm can determine the best classification hyperplane by maximizing the interval, as described in Figure 3
where a dataset with two features (x1 and x2) and two classes (0 and 1) [23], [24]. With the use of support
vector machine technology, data points may be classified by finding a hyperplane in an N-dimensional space.
There are several different hyperplanes for the separation of any two classes of data points. Our goal is to
find a plane that has the most margin. Future data points can be classified more easily by maximizing the
margin distance, which offers some reinforcement. The main flaw with support vector machines is that they
are limited to binary problem classification [25].

Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 
Assessing the effectiveness of data mining tools in classifying and predicting … (Areen Arabiat)
1299
Figure 3. Description of SVM [26]
2.3.4. K-nearest neighbor
KNN classification is a simple data mining technique that forecasts a set's state based on its K
closest neighbors in the training set. It does not require any training but it faced challenges in selecting K
values and performing neighbor searches and distance calculations. KNN is a supervised machine learning
method based on neighbor similarity, classified based on the distance between data points. The primary
obstacles facing KNN are as follows: i) selecting K values and ii) neighbor search and neighbor selection,
encompassing neighbor search and distance calculation [27], [28].
2.4. Data mining tools
Data mining software tools are necessary for both the development and implementation of data
mining techniques. The process of selecting the best tool gets easier as there are more and more options
accessible [29]. The technique of finding important information in large amounts of data is known as data
mining or knowledge mining. It entails several techniques to guarantee that a huge amount of data is
converted into meaningful information, including data translation, cleansing, integration, pattern analysis,
and display [30].
2.4.1. WEKA tool
WEKA, is referred to Waikato Environment for Knowledge Analysis (WEKA), is a machine
learning program developed by Waikato University in New Zealand. It is a Java-based tool that provides
visualization tools and algorithms for predictive modeling and data analysis. It operates on all computing
platforms and includes tasks like data mining, clustering, classification, association, visualization, and feature
selection. The program's user-friendly interface and straightforward settings make it accessible to
inexperienced users [31], [32]. Precision, recall, accuracy, F-measure, MCC, confusion matrix, and other data
may be derived using the WEKA machine learning model to evaluate the result [33]. Figure 4 depicts the
WEKA data mining tool's model.
Figure 4. WEKA data mining tool's model
2.4.2. Orange tool
Orange is a set of machine learning, data mining, and Python scripting tools developed for
interactive data analysis and component-based construction of data mining methods [34]. The bioinformatics
laboratory at the University of Ljubljana has developed the visual data mining program Orange, available for
free and non-commercial download, although primarily designed for instructional purposes, Orange can be
beneficial for data processing and experimental data analysis, offering a platform for experiment selection
[12], [35]. Figure 5 depicts the orange data mining tool's model.

 ISSN: 2502-4752
Indonesian J Elec Eng & Comp Sci, Vol. 34, No. 2, May 2024: 1295-1303
1300
Figure 5. Orange data mining tool's model
3. PERFORMANCE EVALUATION AND CLASSIFICATION
For machine learning tasks like regression and classification, evaluation metrics are essential and
helpful for a variety of tasks. While assuring accurate assessment, proper model evaluation using various
measures can increase predictive parentage and power in addition to avoiding bad predictions when applied
to unknown data. The objective of this procedure is to do a comparative analysis of each classifier and
choose the most accurate one according to the obtained results [36], [37].
3.1. Cross-validation
Since it's easy to use and has a broad range of applications, cross-validation is a common model and
tuning parameter selection technique in statistics and machine learning. Fitting and assessing any potential
model on different data sets is necessary for ensuring accurate assessment. Models are often overfitted by
using conventional techniques like V-fold and leave-one-out. According to preliminary theoretical research,
cross-validation requires a training-testing split ratio of zero to reliably choose the right model under low-
dimensional linear models. Most statistical software programs employ standard split ratios, such as four-to-
one or nine-to-one, because smaller split ratios require larger training samples and lead to less precise model
fitting [38]–[40]. In this research, 10-fold cross-validation is used.
3.2. Confusion matrix
The confusion matrix for the binary classification is written as a 22 matrix. Four measurements
have been published for a confusion matrix: "true positive" (TP), "true negative" (TN), "false positive" (FP),
and "false negative" (FN). The confusion matrix is used to assess classifier performance on datasets in the
multiclass problem. Matrixes to differentiate between the actual and expected values of the model's
constituent parts in Java applications were classified into faulty and non-faulty classes using four confusion
matrix measures: TP, FP, TN, and FN. The figure shows the confusion matrix of binary classification [41]–[44].
Figure 6 shows the confusion matrix of binary classification [45].
Figure 6. Confusion matrix of binary classification [45]

Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 
Assessing the effectiveness of data mining tools in classifying and predicting … (Areen Arabiat)
1301
3.3. Classifiers performance
Datasets and classification algorithms are used to compare the WEKA and Orange data mining tools
in this study. Evaluation criteria include accuracy, sensitivity, precision, and F-measure. By dividing the total
number of properly categorized instances by the total value of instances, the accuracy measure is used to
evaluate performance. Results are assessed utilizing datasets, tools, methods, separation, algorithms, and an
overall total. All tests yielded an 100% categorization accuracy for the research [46]. Table 1 depicts the
classifier's performance.
Table 1. Classifier’s performance [47]
Performance matrices
Equation
Accuracy
TP + TN/TP + FP + TN + FN
Sensitivity
TP/TP + FN
Precision
TP/TP + FP
F-measure
2(Sen  Pre)/ (Sen + Pre)
4. RESULTS AND DISCUSSION
The dataset was classified using a variety of techniques, including SVM, KNN, LR, and RF. Based
on this analysis, the orange tool provided superior results for accuracy (100%) for LR and RF; for KNN and
SVM, the tool achieved CA with values of 99.8% and 99.1%, respectively. On the other hand, the results
using the WEKA tool were also satisfactory. On the other hand, the results using the WEKA tool were also
satisfactory, as SVM obtained a classification accuracy of 99.7%, while KNN, LR, and RF obtained (CA) of
98.7%, 97.6%, and 96.2%, respectively. According to these results, we can notice that the orange data mining
tool is the most effective method for this data, according to the predictions made about traffic congestion.
Table 2 depicts the classifier's performance using the WEKA vs. orange tool, while Figure 7 shows a
comparative analysis of different classifiers' performance. On the other hand, it can also be said that the
Orange3 model proposed in this work for predicting traffic congestion has outperformed previous studies
reported in the literature, such as the study presented by Liu and Wu [8], where the accuracy reached 87.5%
and it was 87.5% in Li et al. [9]. While the accuracy was in the work presented by Majumdar et al. [10], 84–95%.
Table 2. Classifier's performance using WEKA vs. orange tool
RF
LR
SVM
KNN
WEKA
Orange
WEKA
Orange
WEKA
Orange
WEKA
Orange
Accuracy
0.962
1.000
0.976
1.000
0.997
0.991
0.987
0.998
Sensitivity
0.954
1.000
0.975
1.000
0.995
0.991
0.985
0.998
Precision
0.970
1.000
0.977
1.000
0.998
0.991
0.990
0.998
F-measure
0.962
1.000
0.976
1.000
0.997
0.991
0.988
0.998
Figure 7. Comparative analysis of different classifiers' performances
5. CONCLUSION
The purpose of this study was to determine which data mining tool could provide the best prediction
of the accuracy of traffic data. Comparative studies of the tools were conducted to see how successful
different data mining techniques are and how different features affect traffic congestion prediction. The data
was obtained from the Greater Amman Municipality, which mainly contained the traffic volume for the study
area in the capital of the Kingdom of Jordan, Amman, for the year 2018. Various classification methods were
used on the dataset, including RF, LR, KNN, and SVM. Cross-validation is used by 10-fold to improve the

 ISSN: 2502-4752
Indonesian J Elec Eng & Comp Sci, Vol. 34, No. 2, May 2024: 1295-1303
1302
performance of the algorithms. Using the orange tool, RF, and LR have a higher grade of 100% based on this
investigation, while KNN and SVM have scores of 99.8% and 99.1%, respectively. In contrast, using the
WEKA Tool, SVM had a higher grade of 99.7% based on this investigation, while KNN, LR, and RF had
scores of 98.7%, 97.6%, and 96.2, respectively. In the end, we can point out that the results of this
comparison that we presented in our research paper indicate that the model built using the Orange3 data
mining tool outperformed the model built using WEKA, as the accuracy in the first reached 100%.
REFERENCES
[1] T. S. Tamir et al., “Traffic Congestion Prediction using Decision Tree, Logistic Regression and Neural Networks,” IFAC-
PapersOnLine, vol. 53, no. 5, pp. 512–517, 2020, doi: 10.1016/j.ifacol.2021.04.138.
[2] W. Zhuang and Y. Cao, “Short-Term Traffic Flow Prediction Based on a K-Nearest Neighbor and Bidirectional Long Short-Term
Memory Model,” Applied Sciences (Switzerland), vol. 13, no. 4, 2023, doi: 10.3390/app13042681.
[3] Y. Xing, X. Ban, X. Liu, and Q. Shen, “Large-Scale Traffic Congestion Prediction Based on the Symmetric Extreme Learning
Machine Cluster Fast Learning Method,” Symmetry, vol. 11, no. 6, p. 730, May 2019, doi: 10.3390/sym11060730.
[4] T. Adetiloye and A. Awasthi, “Multimodal Big Data Fusion for Traffic Congestion Prediction,” in Multimodal Analytics for Next-
Generation Big Data Technologies and Applications, Cham: Springer International Publishing, 2019, pp. 319–335.
[5] M. W. Ei Leen, N. H. A. Jafry, N. M. Salleh, H. J. Hwang, and N. A. Jalil, “Mitigating Traffic Congestion in Smart and
Sustainable Cities Using Machine Learning: A Review,” Lecture Notes in Computer Science (including subseries Lecture Notes
in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13957 LNCS, pp. 321–331, 2023, doi: 10.1007/978-3-031-
36808-0_21.
[6] T. Saranya, S. Sridevi, C. Deisy, T. D. Chung, and M. K. A. A. Khan, “Performance Analysis of Machine Learning Algorithms in
Intrusion Detection System: A Review,” Procedia Computer Science, vol. 171, pp. 1251–1260, 2020, doi:
10.1016/j.procs.2020.04.133.
[7] S. M. Abdullah et al., " Optimizing Traffic Flow in Smart Cities: Soft GRU-Based Recurrent Neural Networks for Enhanced
Congestion Prediction Using Deep Learning," Sustainability, vol. 15, no. 7, p. 5949, doi: 10.3390/su15075949
[8] Y. Liu and H. Wu, “Prediction of road traffic congestion based on random forest,” in Proceedings - 2017 10th International
Symposium on Computational Intelligence and Design, ISCID 2017, 2018, vol. 2, pp. 361–364, doi: 10.1109/ISCID.2017.216.
[9] Z. Li, P. Liu, C. Xu, H. Duan, and W. Wang, “Reinforcement Learning-Based Variable Speed Limit Control Strategy to Reduce
Traffic Congestion at Freeway Recurrent Bottlenecks,” IEEE Transactions on Intelligent Transportation Systems, vol. 18, no. 11,
pp. 3204–3217, 2017, doi: 10.1109/TITS.2017.2687620.
[10] S. Majumdar, M. M. Subhani, B. Roullier, A. Anjum, and R. Zhu, “Congestion prediction for smart sustainable cities using IoT
and machine learning approaches,” Sustainable Cities and Society, vol. 64, 2021, doi: 10.1016/j.scs.2020.102500.
[11] M. Zahid, Y. Chen, A. Jamal, and M. Q. Memon, “Short term traffic state prediction via hyperparameter optimization based
classifiers,” Sensors (Switzerland), vol. 20, no. 3, 2020, doi: 10.3390/s20030685.
[12] R. Ratra and P. Gulia, “Experimental Evaluation of Open Source Data Mining Tools (WEKA and Orange),” International Journal
of Engineering Trends and Technology, vol. 68, no. 8, pp. 30–35, Aug. 2020, doi: 10.14445/22315381/IJETT-V68I8P206S.
[13] “Greater Amman Municipality.” Jordan, [Online]. Available: http://www.ammancity.gov.jo/en/gam/index.asp (Accessed Sep. 6,
2022).
[14] E. Scornet, G. Biau, and J. P. Vert, “Consistency of random forests,” Annals of Statistics, vol. 43, no. 4, pp. 1716–1741, 2015,
doi: 10.1214/15-AOS1321.
[15] N. Absar et al., “The Efficacy of Machine-Learning-Supported Smart System for Heart Disease Prediction,” Healthcare
(Switzerland), vol. 10, no. 6, 2022, doi: 10.3390/healthcare10061137.
[16] M. Z. Islam, J. Liu, J. Li, L. Liu, and W. Kang, “A semantics aware random forest for text classification,” in International
Conference on Information and Knowledge Management, Proceedings, 2019, pp. 1061–1070, doi: 10.1145/3357384.3357891.
[17] A. Chahal, P. Gulia, N. S. Gill, and J. M. Chatterjee, “Performance Analysis of an Optimized ANN Model to Predict the Stability
of Smart Grid,” Complexity, vol. 2022, 2022, doi: 10.1155/2022/7319010.
[18] F. B. de Santana, W. Borges Neto, and R. J. Poppi, “Random forest as one-class classifier and infrared spectroscopy for food
adulteration detection,” Food Chemistry, vol. 293, pp. 323–332, 2019, doi: 10.1016/j.foodchem.2019.04.073.
[19] H. B. Ly, T. A. Nguyen, and B. T. Pham, “Estimation of Soil Cohesion Using Machine Learning Method: A Random Forest
Approach,” Advances in Civil Engineering, vol. 2021, 2021, doi: 10.1155/2021/8873993.
[20] A. Das, “Logistic Regression,” in Encyclopedia of Quality of Life and Well-Being Research, Cham: Springer International
Publishing, 2021, pp. 1–2.
[21] J. C. Stoltzfus, “Logistic regression: A brief primer,” Academic Emergency Medicine, vol. 18, no. 10, pp. 1099–1104, 2011, doi:
10.1111/j.1553-2712.2011.01185.x.
[22] E. Y. Boateng and D. A. Abaye, “A Review of the Logistic Regression Model with Emphasis on Medical Research,” Journal of
Data Analysis and Information Processing, vol. 07, no. 04, pp. 190–207, 2019, doi: 10.4236/jdaip.2019.74012.
[23] X. Zhang, C. Li, X. Wang, and H. Wu, “A novel fault diagnosis procedure based on improved symplectic geometry mode
decomposition and optimized SVM,” Measurement: Journal of the International Measurement Confederation, vol. 173, 2021,
doi: 10.1016/j.measurement.2020.108644.
[24] B. M. Asl, S. K. Setarehdan, and M. Mohebbi, “Support vector machine-based arrhythmia classification using reduced features of
heart rate variability signal,” Artificial Intelligence in Medicine, vol. 44, no. 1, pp. 51–64, 2008, doi:
10.1016/j.artmed.2008.04.007.
[25] J. Zhou, M. Xiao, Y. Niu, and G. Ji, “Rolling Bearing Fault Diagnosis Based on WGWOA-VMD-SVM,” Sensors, vol. 22, no. 16,
2022, doi: 10.3390/s22166281.
[26] A. Rani, N. Kumar, J. Kumar, and N. K. Sinha, “Machine learning for soil moisture assessment,” Deep Learning for Sustainable
Agriculture, pp. 143–168, 2022, doi: 10.1016/B978-0-323-85214-2.00001-X.
[27] S. Zhang and J. Li, “KNN Classification With One-Step Computation,” IEEE Transactions on Knowledge and Data Engineering,
vol. 35, no. 3, pp. 2711–2723, 2023, doi: 10.1109/TKDE.2021.3119140.
[28] J. Li, J. Zhang, J. Zhang, and S. Zhang, “Quantum KNN Classification With K Value Selection and Neighbor Selection,” IEEE
Transactions on Computer-Aided Design of Integrated Circuits and Systems, 2023, doi: 10.1109/TCAD.2023.3345251.

Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 
Assessing the effectiveness of data mining tools in classifying and predicting … (Areen Arabiat)
1303
[29] R. Mikut and M. Reischl, “Data mining tools,” WIREs Data Mining and Knowledge Discovery, vol. 1, no. 5, pp. 431–443, Sep.
2011, doi: 10.1002/widm.24.
[30] S. Verma and P. Rattan, “Introduction To Data Mining Tools and Techniques & Applications: a Review,” in Business, no. July,
2021, [Online]. Available: https://www.researchgate.net/profile/Rashmi-Gujrati-2/publication/355170587_Role_of_Technology
_in_New_Decades/links/6163de531eb5da761e794894/Role-of-Technology-in-New-Decades.pdf#page=57.
[31] Sunita B Aher and LOBO L.M.R.J., “Data Mining in Educational System using WEKA,” 2011, [Online]. Available:
http://www.ijcaonline.org/icett/number3/icett021.pdf.
[32] E. Kulkarni G. and R. Kulkarni B., “WEKA Powerful Tool in Data Mining,” International Journal of Computer Applications
National Seminar on Recent Trends in Data Mining, vol. 5, no. Rtdm, pp. 975–8887, 2016, [Online]. Available:
http://research.ijcaonline.org/rtdm2016/number2/rtdm2575.pdf.
[33] P. Debnath et al., “Analysis of Earthquake Forecasting in India Using Supervised Machine Learning Classifiers,” Sustainability,
vol. 13, no. 2, p. 971, Jan. 2021, doi: 10.3390/su13020971.
[34] J. Demšar et al., “Orange: Data mining toolbox in python,” Journal of Machine Learning Research, vol. 14, pp. 2349–2353, 2013.
[35] Z. Dobesova, “Experiment in Finding Look-Alike European Cities Using Urban Atlas Data,” ISPRS International Journal of Geo-
Information, vol. 9, no. 6, p. 406, Jun. 2020, doi: 10.3390/ijgi9060406.
[36] Ž. Vujović, “Classification Model Evaluation Metrics,” International Journal of Advanced Computer Science and Applications,
vol. 12, no. 6, pp. 599–606, 2021, doi: 10.14569/IJACSA.2021.0120670.
[37] M. R. Mahmood, M. B. Abdulrazzaq, S. R. M. Zeebaree, A. K. Ibrahim, R. R. Zebari, and H. I. Dino, “Classification techniques’
performance evaluation for facial expression recognition,” Indonesian Journal of Electrical Engineering and Computer Science,
vol. 21, no. 2, pp. 1176–1184, 2020, doi: 10.11591/ijeecs.v21.i2.pp1176-1184.
[38] P. Zhang, “Model Selection Via Multifold Cross Validation,” The Annals of Statistics, vol. 21, no. 1, 2007, doi:
10.1214/aos/1176349027.
[39] J. Lei, “Cross-Validation With Confidence,” Journal of the American Statistical Association, vol. 115, no. 532, pp. 1978–1997,
2020, doi: 10.1080/01621459.2019.1672556.
[40] J. Shao, “Linear Model Selection by Cross-validation,” Journal of the American Statistical Association, vol. 88, no. 422, pp. 486–
494, Jun. 1993, doi: 10.1080/01621459.1993.10476299.
[41] R. Rajalakshmi and C. Aravindan, “A Naive Bayes approach for URL classification with supervised feature selection and
rejection framework,” Computational Intelligence, vol. 34, no. 1, pp. 363–396, 2018, doi: 10.1111/coin.12158.
[42] J. Font, L. Arcega, Ø. Haugen, and C. Cetina, “Achieving feature location in families of models through the use of search-based
software engineering,” IEEE Transactions on Evolutionary Computation, vol. 22, no. 3, pp. 363–377, 2018, doi:
10.1109/TEVC.2017.2751100.
[43] Y. S. Taspinar, M. Koklu, and M. Altin, “Classification of flame extinction based on acoustic oscillations using artificial
intelligence methods,” Case Studies in Thermal Engineering, vol. 28, 2021, doi: 10.1016/j.csite.2021.101561.
[44] A. Feyzioglu and Y. S. Taspinar, “Beef Quality Classification with Reduced E-Nose Data Features According to Beef Cut
Types,” Sensors, vol. 23, no. 4, 2023, doi: 10.3390/s23042222.
[45] I. Markoulidakis, G. Kopsiaftis, I. Rallis, and I. Georgoulas, “Multi-Class Confusion Matrix Reduction method and its application
on Net Promoter Score classification problem,” ACM International Conference Proceeding Series, pp. 412–419, 2021, doi:
10.1145/3453892.3461323.
[46] R. Panigrahi et al., “Performance assessment of supervised classifiers for designing intrusion detection systems: A comprehensive
review and recommendations for future research,” Mathematics, vol. 9, no. 6, 2021, doi: 10.3390/math9060690.
[47] F. Sajid et al., “Secure and Efficient Data Storage Operations by Using Intelligent Classification Technique and RSA Algorithm
in IoT-Based Cloud Computing,” Scientific Programming, vol. 2022, 2022, doi: 10.1155/2022/2195646.
BIOGRAPHIES OF AUTHORS
Areen Arabiat obtained B.Sc in computer engineering in 2005 from Al Balqa
Applied University (BAU), and her MSc in Intelligent Transportation Systems (ITS) from Al
Ahliyya Amman University (AAU) in 2022. She is currently a computer lab supervisor at the
Faculty of Engineering/Al-Ahliyya Amman University (AAU) since 2013. Her research
interests are focused on the following areas: machine learning, data mining, artificial
intelligence, and image processing. She can be contacted email: a.arabiat@ammanu.edu.jo.
Muneera Altayeb obtained a bachelor’s degree in computer engineering in 2007,
and a master’s degree in communications engineering from the University of Jordan in 2010.
She has been working as a lecturer in the Department of Communications and Computer
Engineering at Al-Ahliyya Amman University since 2015, in addition to her administrative
experience as assistant dean of the Faculty of Engineering during the period (2020-2023).
Her research interests focus on the following areas: digital signals and image processing,
machine learning, robotics, and artificial intelligence. She can be contacted at email:
m.altayeb@ammanu.edu.jo.