IAES International Journal of Artificial Intelligence (IJ-AI)

Vol. 11, No. 4, December 2022, pp. 1535~1544

ISSN: 2252-8938, DOI: 10.11591/ ijai.v 11.i4.pp1535-1544  1535

Weather prediction performance evaluation on selected

machine learning algorithms

Muyideen AbdulRaheem1 , Joseph Bamidele Awotunde1,3,4 , Abidemi Emmanuel Adeniyi 2,3,4,5 ,

Idowu Dauda Oladipo1 , Sekinat Olaide Adekola1
1
Department of Computer Science, University of Ilorin, Ilorin, Nigeria
2
Department of Computer Science, Landmark University, Omu-Aran, Nigeria
3
Industry, Innovation and Infrastructure Research Group, Landmark University, Omu-Aran, Nigeria, SDG 9
4
Climate Action Research Group, Landmark University, Omu-Aran, Nigeria, SDG 13
5
Life on Land Research Group, Landmark University, Omu-Aran, Nigeria, SDG 15

Article Info ABSTRACT

Article history: Prediction of weather has been proved useful in the early warning on the
impacts of weather on several areas of human livelihood. For example, the
Received Jul 27, 2021 provision of decisions for autonomous transportation to reduce traffic
Revised May 31, 2022 congestion and accidents during the rainy season. However, providing the
Accepted Jun 29, 2022 most accurate and effective forecasting model for weather forecasts has been
a challenge. Hence, machine learning (M L) techniques and factors
influencing weather prediction need to be investigated. Data scientists are
Keywords: yet to discover the best models for weather prediction. Therefore, this study
compares three M L classification techniques for weather prediction. A web-
Data mining based software application was developed using Flask App to demonstrate
Decision trees weather modeling using three M L models, and the data used for the study
K-nearest neighbors was obtained from Kaggle. For the weather prediction; a decision tree (DT),
Logistic regression k-nearest neighbor (k-NN), and logistic regression (LR) classifier method
Normalization were suggested, and comparisons were made between the three
Weather forecasting classifications techniques. The accuracy results show that with a 100%
accuracy rate, the DL classifier outperforms the k-NN with a 78% accuracy
rate and LR with a 93% accuracy rate. The results show that the application
of M L models gives accurate results on weather prediction.
This is an open access article under the CC BY-SA license.

Corresponding Author:
Abidemi Emmanuel Adeniyi
Department of Computer Science, Landmark University
Omu-Aran, Kwara State, Nigeria
Email: [email protected]

1. INTRODUCTION
Nowadays, data mining is playing a vital role in weather prediction and climate change studies and
it is one of the most important fields of research to extract useful knowledge from large datasets [1], [2].
Weather forecasting and climate change studies are significant research areas that aid in the prediction of
useful information from weather data to enhance weather usability, better understanding, and more reliable
prediction of weather-related activities. Data mining, also known as knowledge discovery, is becoming
increasingly important because it aids in the analysis of data from various perspectives and the
summarization of that data into useful information [3], [4]. Every data scientist is faced with the following
fundamental questions: which predictive model is more appropriate for the problem at hand? What typ e of
programming language to use? And what type of tools can be used for efficient outcomes. Python has several
built-in machine learning libraries that provide a strong computational capability in data mining.

Journal homepage: http://ijai.iaescore.com

1536  ISSN: 2252-8938

Weather forecasting is the use of science and technology to predict atmospheric conditions for a
specific location and at a specific time. For centuries, people have attempted to predict the weather
informally, and formally since the 19th century [5], [6]. Weather predictions are used by a wide range of
users such as agriculture [7], [8]. Weather warnings are important predictions because they protect people
and property. Forecasts dependent on temperature and precipitation, for example, are important for
agriculture as well as commodity traders. People use weather forecasts daily to decide what to wear. Since
heavy rain, high temperatures, snow, and wind chill greatly limit outdoor activities, forecasts can be used to
schedule activities around these events, as well as to prepare for and withstand them. The reliance on weather
forecasts by farmers, traders, transportation industries, and individual call for an accurate weather forecast.
Data mining is a technique for training a computer on how to make data -driven decisions. This decision
could be as simple as predicting the weather for tomorrow, blocking a spam email from acce ssing your
inbox, detecting the language of a website, or discovering a new relationship on a dating website. Data
mining has a wide range of applications, with new ones appearing regularly. Moreover, prediction is used to
foretell the next event based on the current state of events. In intelligence environments, prediction is
important because it detects the repetitive trend and predicts what will happen in the future. Modern farmers
and business people need accurate real-time prediction tools to forecast the weather and climate change.
Data mining is widely used in weather and climate change studies to accurately forecast the weather.
Weather forecasting applications tend to be complex to apply, inaccurate and unreliable as a result of the
methods used in the application. Accurate weather prediction requires correct weather parameters, an
efficient data mining approach, and apparatus. Weather prediction has become one of the most challenging
scientific and technological problems in the last century all over the world [9]. The inaccuracy always calls
for an alternative approach to address the issue because accurate weather prediction requires correct weather
parameters, an efficient data mining approach, and apparatus. However, the proposed system employs three
supervised learning algorithms, including logistic regression (LR), k-nearest neighbors (k-NN), and decision
tree (DT) classification, with Spyder as the IDE adopting Python as the programming language to determine
the most suitable technique for weather prediction. Data mining techniques may be used for weather
forecasting and climate change studies if there is enough case data. Weather forecasting has historically bee n
conducted traditionally with disturbances and uncertainty when calculating the initial atmospheric conditions.
Table 1 provided the summary of the related works in weather and rainfall prediction.

Table 1. Related works

Authors Research T opics T echniques/Algorithms T ools
Anton et al. [7] Collaborative data mining in k-nearest neighbor model (k-NN) T he data analysis was
agriculture for prediction of soil local polynomial regression (LPR) conducted with the aid of a
moisture and temperature neural net model (NN) and support SQL command and a Microsoft
vector machine (SVM) Access database.
Shah et al. [10] Rainfall prediction: accuracy Neural network A text tool was used (MS Word
enhancement using machine Decision tree as the formatting tool)
learning and forecasting ARIMA
techniques
Zaman [11] Machine learning model on Classification algorithms (Naive Apache Spark was used for the
rainfall - a predicted approach for Bayes, random forest classifier, machine learning library and
Bangladesh and decision tree algorithm) scala as a development
Regression algorithm (linear platform.
regression, random forest
regression)
Shivang and Weather Prediction for Indian Linear regression Python (manual programming)
Sridhar [12] location using machine learning Functional regression
Neural network
PosPieszny [13] Application of data mining C4.5 tree Not indicated
techniques in project Random tree (RT)
management-an overview Classification and regression tree
(CART )
T alib et al. [14] Application of data mining K-means clustering algorithm Waikato Environment for
techniques in weather data Decision tree algorithm Knowledge Analysis (WEKA)
analysis data mining tools were used
Gouda and Data mining for weather and Classification Not indicated
Chandrika [15] climate studies Clustering
Joshi et al. [9] Weather forecasting and climate Decision tree classifiers Both decision trees and
changing using data mining decision tree rules were created
application using the See5 program.
Olaiya and Application of data mining Artificial neural network T he C5 Decision T ree classifier
Adeyemo [16] techniques in weather prediction Decision tree algorithms algorithm has been
and climate change Studies implemented in See5.

Int J Artif Intell, Vol. 11, No. 4, December 2022: 1535-1544

Int J Artif Intell ISSN: 2252-8938  1537

Table 1. Related works (continue)

Authors Research T opics T echniques/ Algorithms T ools
Cortez and Morais [17] A Data mining approach to Multiple regression (MR) T he open-source library Rainer
predict forest fires using Decision trees (DT) and (for the R statistical
meteorological data Random forests (RF) environment) was used.
Neural networks (NN)
Support vector machines
(SVM)
Oladipo et al. [18] Prediction and analysis of Classification WEKA tool was used as the
student performance by data Association software for data mining
mining in WEKA

Many other studies have contributed to the production of weather prediction systems that have
yielded significant results and achieved the goals for which they were created; each study focuses on a
specific challenge and investigates solutions from a uniqu e perspective [19]–[21]. However, these techniques
exhibit several defects such as; they may not completely experience from the existing weather data, and may
not have the most accurate results. This study aims to compare the performance of three different
classification machine learning algorithms in predicting weather and climate change for a specific location
using Python because of its speed and libraries
Predictive modeling has gone through a revolution in recent years, thanks to advances in computer
computing power [22]–[24]. Thousands of models can now be run on multiple cores at high GHz speeds,
making predictive modeling more effective and affordable than ever b efore [25]. However, every data
scientist is faced with the following fundamental questions: which predictive model is more appropriate for
the problem at hand? What type of programming language to use? And what type of tools can be used for
efficient outcomes. Since Python has several built-in machine learning libraries that provide a strong
computational capability in data mining, this research contribution will be an adequate, efficient, and reliable
resource for weather prediction applications.

2. RESEARCH METHOD
In this paper, the methodology used various classifiers to forecast weather. The daily meteorological
data consisting of DATE, PRCP (which is the precipitation for that day), TMAX (which is the maximum
temperature for that day), TMIN (which is the minimum temperature for that day), and RAIN of Seattle-
Tacoma International Airport will be trained using Python inbuilt library, Scikit -learn library which consists
of DT, LR and K-NN machine learning algorithms. These three machine learning algorithms will be used in
this study to conduct classification rules and provide a model for weather forecasting. The target variable in
this study is the Rainfall which is 1 (TRUE) if rain was observed on that day and 0 (FALSE) if it was not.
Inconsistency and unreliability of meteorological factors make this research focus on the use of Python as the
tool for analyzing the weather data to develop approaches that can recognize the best technique that can
handle such an inconsistent pattern effectively. The main problem of this research work is to test the ability
of some classification machine learning algorithms to predict weather data and to analyze, based on the data
used, the accuracy of the machine learning algorithms.

2.1. Architecture of research methodolog y

Machine learning, also known as predictive modeling, is mainly concerned with reducing a model's
error, and also, the most precise predictions are expected. Algorithm tuning is one of the most popular
approaches to improving the accuracy of an algorithm. Mac hine learning algorithms are known to be
influenced by parameters. These variables have a major impact on the learning process' outcome. Parameter
tuning aims to find the best value for each parameter to improve the model's accuracy. To tune these
parameters, you'll need to know what they mean and how they affect the model individually. To understand
the main problem of the research work, it can be considered in the following two steps as shown in Figure 1.

2.1.1. Step 1
Weather data consists of complete records of daily rainfall pattern from January 1st, 1978 to December
12, 2017 (69 years) and constitutes parameters (25,551 rows x 5 columns) namely, DAY, MONTH, PRCP
(inch), TMAX °F, TMIN °F, RAIN (TRUE or FALSE). The main problem of this part is to do da ta wrangling
on the weather data. The variables have to be chosen in such a way that it includes the effect of rainfall. Then
the problem comes to decide the weather variables which have a significant impact on the output variable. We
need to decide various inputs and output parameters and check for anomalies using various Pandas library tools
to check for missing values, performing data wrangling, and then process the data.

Weather prediction performance evaluation on selected … (Muyideen AbdulRaheem)

1538  ISSN: 2252-8938

Weather Prediction Problem

Step 1 Step 2

Explore Data DecisionTreeClassifier( )

KNeighborsClassifier( )
 Normalization LogisticRegression( )
 Select data
 Split data

Train Predict

Examine Results

Figure 1. Methodology process

These steps are as: i) Import data: Using Pandas to import the data set, which is a two-dimensional
structure in an arrangement of rows and columns ; ii) Data normalization: This involves tasks such as
removing and/or modifying duplicate data; iii) Data splitting: To split the data into training sets and testing
sets using a higher percentage for the training sets ; and iv)Data selection: This involves the process of
dividing and selecting the dataset into two parts X and y, which are variables data and feature data
respectively.

2.1.2. Step 2
The authors need to develop the model using the machine learning algorithms
(DecisionTreeClassifier(), KneighborClassifier(), and LogisticRegression()) to train data and predict the
output value of rain. The same data is used for accuracy check and error is calculate d between the predicted
and observed value of rain. Lastly, error evaluation and analysis visualization are done on each model to
examine the results given. These steps are as: i) Data modelling: applying the machine learning algorithms
which are DecisionTreeClassifier(), KneighborsClassifier() and LogisticRegression() that are provided in
Scikit-learn library; ii) Model evaluation: This is the stage whereby the models are tested to measure their
accuracy using error evaluation metrics such as accuracy score, and confusion matrix.

2.2 Data collection

The data collected contains the 49 years of weather data, from Seattle-Tacoma International Airport.
It contains for each day the minimum temperature, maximum temperature, precipitation, and rainfall. The
temperature is measured in Fahrenheit and precipitation is measured in inches. Rain contains Boolean values.
The data is already in the .csv file. Table 2 shows the sample from the dataset.
In this study, to perform prediction, the weather dataset is split into two sets, input and output
variables. Input data is loaded as X vector whose columns are DAY, MONTH, PRCP, TMAX, and TMIN,
these are the independent variables and the target is loaded as y, which is the dependent variable co ntaining
the data of Rainfall (Boolean) as well-defined in Table 3. To predict the weather, we need to characterize the
parameters which indirectly affect the rainfall. The variables of weather prediction are maximum temperature
(TMAX o F) and minimum temperature (TMIM o F).
Figure 2 displayed the proposed model framework used for weather prediction. The proposed model
was use for the prediction of weather using selected machine learning algorithms, the framework consists of
the following: i) Pre-processing, ii) normalization, iii) prediction, and iv) evaluation of the proposed system
using various metrics.

Int J Artif Intell, Vol. 11, No. 4, December 2022: 1535-1544

Int J Artif Intell ISSN: 2252-8938  1539

Table 2. Data sample

DAT E DAY MONT H PRCP T MAX T MIN RAIN
01/01/2017 1 1 0.43 37 28 T RUE
02/01/2017 2 1 0 34 26 FALSE
03/01/2017 3 1 0 33 21 FALSE
04/01/2017 4 1 0 36 22 FALSE
05/01/2017 5 1 0 35 21 FALSE
06/01/2017 6 1 0 40 20 FALSE
07/01/2017 7 1 0 37 29 FALSE
08/01/2017 8 1 0.45 45 35 T RUE
09/01/2017 9 1 0.05 42 34 T RUE
10/01/2017 10 1 0.07 40 32 T RUE

Table 3. List of input and output variables

Variable Weather Parameter T ype
X1 DAY (int) Input
X2 MONT H (int) Input
X3 PRCP (float) Input
X4 T MAX (int) Input
X5 T MIN (int) Input
y1 Rainfall (int) T arget

Figure 2. Framework model for weather forecast

Many researchers globally have been working on weather prediction using historical data and
machine learning models in recent years to prevent weather prediction unreliability [26]–[28]. In this article,
the study used classification algorithms for the weather prediction model to classify data for the target
column used to predict rain in Seattle weather data. If the qualified model is for predicting any of two target
classes, the main aim of the classification algorithm is to predict the target class (Yes/No). When used
correctly, it will assist in predicting how a variable will appear in the future based on other variables. Spyder
Weather prediction performance evaluation on selected … (Muyideen AbdulRaheem)
1540  ISSN: 2252-8938

IDE was used for data collection, model building, preparation, testing, and prediction. I had introduced
various types of functionalities for common tasks during this process. The data are trained, tested, and saved
the model to disk for live data testing in the future. Testing for those models is going to be difficult and
inconvenient. As a result, the study decided to create a web application that would allow anyone to use the
developed model to predict rainfall using a web browser.

3. RESULTS AND DISCUSSION

Building an efficient machine learning program necessitates evaluating a model. Model output is
explained using evaluation criteria, which often aid in distinguishing between model outcomes. The study
used a confusion matrix as well as four evaluation metrics: accuracy score, precision score, recall sc ore, and
F1-score, which are the most commonly used evaluation metrics for machine learning models. The following
are the model evaluation and outcomes. This experiment's decision tree classifier resultant tree is shown in
Figure 3. The top of the tree (known as the root node) has arrows pointing to and away from it, while
intermediate nodes have arrows pointing to and away from them. Leaves are nodes at the bottom of a tree
that do not have any edges pointing away from them.

Figure 3. Decision tree classification

The experiment's decision tree classifier confusion matrix results are shown in Figure 4(a). The true
positive (TP) of the entire test set in this decision tree classifier confusion matrix was 95.37% of the dataset
belongs to the positive class, whereas 4.63% of data points belong to the negative category. The false positive
(FP) and false negative (FN) rates are both 0%, indicating that the decision tree model did not classify any
data incorrectly. The experiment's k-nearest neighbor confusion matrix results shown in Figure 4(b) reveals
that the TP of the entire test set in this k-Neighbor classifier confusion matrix was 46.68%, indicating that the
model correctly identified 46.68% of positive class data points. In addition, the true negative (TN) was
31.42%, indicating that the model correctly identified 31.42% of negative class data points. The FP was
12.21%, which means the model incorrectly classified 12.21% of negative class data points as positive, and
the FN was 9.68%, which means the model incorrectly classified 9.68% of p ositive class points as negative.
The experiment's logistic regression confusion matrix results shown in Figure 4(c). The TP of the
entire test set in this logistic regression classifier confusion matrix was 56.00%, indicating that the model
correctly identified 56.00% of positive class data points. Furthermore, the TN was 36.59%, indicating that the
model correctly identified 36.59% of negative class data points. The FP rate was 7.04%, indicating that the
model incorrectly classified 7.04% of negative clas s data points as positive, and the FN rate was 19%,
indicating that the model incorrectly classified 19% of positive class points as negative.

Int J Artif Intell, Vol. 11, No. 4, December 2022: 1535-1544

Int J Artif Intell ISSN: 2252-8938  1541

(a) (b)

(c)

Figure 4. Comparing the confusion matrix results for weather prediction in (a) CART, (b) K-NN, and (c) LR
machine learning

The K-NN algorithm predicts the values of new data points based on ‘feature similarity.' This
implies that a value is given to the new point based on how similar it is to the points in the training set. It
effectively measures the difference between a new data point and the training set's previous data points. It
places the data point in the class that contains the majority of the K data points. The resultant classification
model for this experiment k-nearest neighbor classification illustration is shown in Figure 5(a), and the
logistic regression is displayed in Figure 5(b).
The dependent variable in logistic regression is a binary variable that contains data coded as 1 (yes,
performance, etc.) or 0 (no, failure, etc.). To put it another way, the logistic regression model predicts P(Y=1)
as a function of X and outputs a constant value. The TMAX represents the x-axis in this logistic regression
graph, while the TMIN represents the y-axis. The blue circles represent observations that are categorized as
zeros, indicating that rain is not occurring, while the orange circles represent observations that are classified
as ones, indicating that rain is occurring. Table 4 displayed the model evaluation results using various metrics
like accuracy, precision, recall, and F1-score.

Table 4. Model evaluation (evaluation metrics)

MODELS Accuracy Precision Recall F1-Score
Decision T ree 100 1.0 1.0 1.0
k-Nearest Neighbor (k-NN) 78.10 0.76 0.72 0.74
Logistic Regression (LR) 93 0.98 0.83 0.90

The models were measured using four different evaluation metrics. The evaluation metrics allow for
a quick assessment of the model's performance. To demonstrate how the models worked, the accuracy score,
precision score, recall score, and F1-score evaluation metrics were used in this experiment. The results
showed that the DT classifier performs better with an accuracy of 100%, and 100% all through other metrics
used. According to the evaluation metrics, the DT classifier has the highest evaluation scores, followed by
LR, and k-NN classifier.

Weather prediction performance evaluation on selected … (Muyideen AbdulRaheem)

1542  ISSN: 2252-8938

(a)

(b)

Figure 5. Comparing scatter graph results for weather prediction in (a) k-nearest neighbor and (b) logistic
regression

4. CONCLUSION
Improving a model's productivity can be challenging at times, and a predictive model can be built in
various ways since there are no set rules to follow. The production of huge data from weather generated
datasets have helped the prediction of weather forecast using machine learning (ML)-models. This has helped
in the "proof to speak for itself" rather than relying on assumptions and bad associations. Since variables like
humidity and wind speed influence the weather, providing more data leads to better and more accurate
models. This intuition develops over time as a result of practice and experience. This is compounded by the
fact that certain algorithms are better suited to some types of datasets than others. therefore, this paper
predicts weather forecasts using three classifiers using the Seattle weather dataset to test the effectiveness of
these three models. These models are validated using Kaggle's meteorological data, which includes, the date,
maximum temperature, minimum temperature, precipitation, and rain. The decision tree algorithm, which has
a 100% accuracy rate, outperforms the logistic regression (which has a 93% accuracy rate), and the k-nearest

Int J Artif Intell, Vol. 11, No. 4, December 2022: 1535-1544

Int J Artif Intell ISSN: 2252-8938  1543

neighbor algorithm (which has a 78% accuracy rate). The study achieved an appropriate label of accu racy for
the decision tree algorithm in terms of rain prediction. Future work can make use of the open works up by
training the weather datasets using intuitive optimization on the datasets parameter this will result in better
and more accurate models.

REFERENCES
[1] H. A. Issad, R. Aoudjit, and J. J. P. C. Rodrigues, “A comprehensive review of Data Mining techniques in smart agriculture,”
Eng. Agric. Environ. Food, vol. 12, no. 4, pp. 511–525, Oct. 2019, doi: 10.1016/j.eaef.2019.11.003.
[2] K. Soomro, M. N. M. Bhutta, Z. Khan, and M. A. Tahir, “Smart city big data analytics: An advanced review,” Wiley Interdiscip.
Rev. Data Min. Knowl. Discov., vol. 9, no. 5, p. e1319, Jun. 2019, doi: 10.1002/widm.1319.
[3] Z. Ge, Z. Song, S. X. Ding, and B. Huang, “Data mining and analytics in the process industry: T he role of machine learning,”
IEEE Access, vol. 5, pp. 20590–20616, 2017, doi: 10.1109/access.2017.2756872.
[4] A. M. Abdu, M. M. M. Mokji, and U. U. U. Sheikh, “Machine learning for plant disease detection: an investigative comparison
between support vector machine and deep learning,” IAES Int. J. Artif. Intell., vol. 9, no. 4, p. 670, Dec. 2020, doi:
10.11591/ijai.v9.i4.pp670-683.
[5] P. Steer, “T he climates of the victorian novel: seasonality, weather, and regional fiction in Britain and Australia,”
PMLA/Publications Mod. Lang. Assoc. Am., vol. 136, no. 3, pp. 370–385, May 2021, doi: 10.1632/s0030812921000286.
[6] I. Loor and J. Evans, “Understanding the value and vulnerability of informal infrastructures: Footpaths in Quito,” J. Transp.
Geogr., vol. 94, p. 103112, Jun. 2021, doi: 10.1016/j.jtrangeo.2021.103112.
[7] C. A. Anton, O. Matei, and A. Avram, “Collaborative data mining in agriculture for prediction of soil moisture and temperature,”
in Computer Science On-Line Conference, Springer International Publishing, 2019, pp. 141 –151.
[8] B. Das, B. Nair, V. K. Reddy, and P. Venkatesh, “Evaluation of multiple linear, neural network and penalised regression model s
for prediction of rice yield based on weather parameters for west coast of India,” Int. J. Biometeorol., vol. 62, no. 10, pp. 1809–
1822, Jul. 2018, doi: 10.1007/s00484-018-1583-6.
[9] A. Joshi, B. Kamble, V. Joshi, K. Kajale, and N. Dhange, “Weather forecasting and climate changing using data mining
application,” Int. J. Adv. Res. Comput. Commun. Eng., vol. 4, no. 3, pp. 19–21, Mar. 2015, doi: 10.17148/ijarcce.2015.4305.
[10] U. Shah, S. Garg, N. Sisodiya, N. Dube, and S. Sharma, “Rainfall prediction: accuracy enhancement using machine learning and
forecasting techniques,” in 2018 Fifth International Conference on Parallel, Distributed and Grid Computing (PDGC), Dec.
2018, pp. 776–782, doi: 10.1109/pdgc.2018.8745763.
[11] Y. Zaman, “Machine learning model on rainfall-a predicted approach for Bangladesh,” United International University, 2018.
[12] J. Shivang and S. S. Sridhar, “Weather prediction for indian location using Machine learning,” Int. J. Pure Appl. Math., vol. 118,
no. 22, pp. 1945–1949, 2018.
[13] P. PosPieszny, “Application of data mining techniques in project management–an overview,” CEA Ann., vol. 43, pp. 199–220,
2017, [Online]. Available: http://rocznikikae.sgh.waw.pl/p/roczniki_kae_z43_12.pdf.
[14] M. R. T alib, T . Ullah, M. U. Sarwar, M. K. Hanif, and N. Ayub, “Application of data mining techniques in weather data
analysis,” Int. J. Comput. Sci. Netw. Secur., vol. 17, no. 6, pp. 22–28, 2017, [Online]. Available:
http://paper.ijcsns.org/07_book/201706/20170604.pdf.
[15] K. C. Gouda and and Chandrika M, “Data mining for weather and climate studies,” Int. J. Eng. Trends Technol., vol. 32, no. 1,
pp. 29–32, Feb. 2016, doi: 10.14445/22315381/ijett -v32p206.
[16] F. Olaiya and A. B. Adeyemo, “Application of data mining techniques in weather prediction and climate change studies,” Int. J.
Inf. Eng. Electron. Bus., vol. 4, no. 1, pp. 51–59, Feb. 2012, doi: 10.5815/ijieeb.2012.01.07.
[17] P. Cortez and A. de J. R. Morais, “A data mining approach to predict forest fires using meteorological data,” Environmental
Science, Computer Science. Sep. 2007, [Online]. Available: http://www3.dsi.uminho.pt/pcortez/fires.pdf.
[18] I. D. Oladipo et al., “An improved course recommendation system based on historical grade data using logistic regression,” in
Communications in Computer and Information Science, Springer International Publishing, 2021, pp. 207 –221.
[19] D. F. Sengkey, A. Jacobus, and F. J. Manoppo, “Effects of kernels and the proportion of training data on the accuracy of SVM
sentiment analysis in lecturer evaluation,” IAES Int. J. Artif. Intell., vol. 9, no. 4, p. 734, Dec. 2020, doi:
10.11591/ijai.v9.i4.pp734-743.
[20] H. Hertina et al., “Data mining applied about polygamy using sentiment analysis on twitters in Indonesian perception,” Bull.
Electr. Eng. Informatics, vol. 10, no. 4, pp. 2231–2236, Aug. 2021, doi: 10.11591/eei.v10i4.2325.
[21] N. Razali, S. Ismail, and A. Mustapha, “Machine learning approach for flood risks prediction,” IAES Int. J. Artif. Intell., vol. 9,
no. 1, p. 73, Mar. 2020, doi: 10.11591/ijai.v9.i1.pp73 -80.
[22] M. A. Febriantono, S. H. Pramono, R. Rahmadwati, and G. Nagh dy, “Classification of multiclass imbalanced data using cost -
sensitive decision tree C5.0,” IAES Int. J. Artif. Intell., vol. 9, no. 1, p. 65, Mar. 2020, doi: 10.11591/ijai.v9.i1.pp65 -72.
[23] M. Y. I. Basheer, S. Mutalib, N. H. A. Hamid, S. Abdul-Rahman, and A. M. A. Malik, “Predictive analytics of university student
intake using supervised methods,” IAES Int. J. Artif. Intell., vol. 8, no. 4, p. 367, Dec. 2019, doi: 10.11591/ijai.v8.i4.pp367 -374.
[24] J. B. Awotunde, C. Chakraborty, and A. E. Adeniyi, “Intrusion detection in industrial internet of things network-based on deep
learning model with rule-based feature selection,” Wirel. Commun. Mob. Comput., vol. 2021, Sep. 2021, doi:
10.1155/2021/7154587.
[25] L. T iwari et al., “Detection of lung nodule and cancer using novel Mask-3 FCM and TWEDLNN algorithms,” Measurement, vol.
172, p. 108882, Feb. 2021, doi: 10.1016/j.measurement.2020.108882.
[26] J. B. Awotunde, S. O. Folorunso, A. K. Bhoi, P. O. Adebayo, and M. F. Ijaz, “Disease diagnosis system for IoT -based wearable
body sensors with machine learning algorithm,” in Intelligent Systems Reference Library, vol. 209, Springer Singapore, 2021, pp.
201–222.
[27] S. Susanto, D. D. Budiarjo, A. Hendrawan, and P. T . Pungkasanti, “T he implementation of intelligent systems in automating
vehicle detection on the road,” IAES Int. J. Artif. Intell., vol. 10, no. 3, p. 571, Sep. 2021, doi: 10.11591/ijai.v10.i3.pp571 -575.
[28] X. Ren et al., “Deep learning-based weather prediction: A survey,” Big Data Res., vol. 23, p. 100178, Feb. 2021, doi:
10.1016/j.bdr.2020.100178.

Weather prediction performance evaluation on selected … (Muyideen AbdulRaheem)

1544  ISSN: 2252-8938

BIOGRAPHIES OF AUTHORS

Dr. Muyideen AbdulRaheem is a Lecturer in the Department of Computer Science,

University of Ilorin, Ilorin, Nigeria. He obtained a Bachelor of Technology Computer Science at
the Prestigious Abubakar Tafawa Balewa University Bauchi, in 1997, M aster of Science in
M athematics (Computer Option) and Ph.D. Computer Science at the University of Ilorin, Ilorin,
Nigeria. His research area is on computer security, grid computing security and data and
information security. He can be contacted at: email: [email protected].

Joseph Bamidele Awotunde received his B.Sc. degree in M athematics/ Computer

Science from Federal University of Technology, M inna, Nigeria, in 2007. M .Sc. and Ph.D.
degrees in Computer Science from the University of Ilorin, Ilorin, Nigeria, in 2014 and 2019
respectively. From 2012 to 2015, and 2018, he was a Computer Science Instructor with the
University School, University of Ilorin, Ilorin, Nigeria. From 2017 to 2018, he was a Lecturer II
with the M cPherson University, Ijebo, Seriki-Sotayo, Nigeria. Since 2019, he has been a Lecturer
II with the Computer Science Department, University, of Ilorin, Ilorin, Nigeria. He is the author of
more than 40 articles, 30 chapter in book, and more than 25 Conference Proceedings. His research
interests include Information Security, Cybersecurity, Bioinformatics Artificial Intelligence,
Internet of M edical Things, Wireless Body Sensor Networks, Wireless Networks, Telemedicine,
m-Health/e-health, and M edical Imaging, Information Security, Software Engineering, and
Biometrics. He is a member of International Association of Engineers and Computer Scientist
(M IAENG), Computer Professional Registration Council of Nigeria (M CPN), Nigeria Computer
Society (M NCS), and Internet Society. He can be contacted at email:
[email protected], [email protected].

Abidemi Emmanuel Adeniyi is a Staff in the Department of Computer Science,

College of Pure and Applied Sciences, Landmark University, Omu-Aran, Nigeria. He received
Bachelor of Science in Computer Science from Ladoke Akintola University of Technology,
Ogbomosho. He also begged M asters of Science in Computer Science from Landmark University,
Omu-Aran. His area of research includes computer security, Computational Complexity Analysis,
Cryptographic Algorithms and Information Security. He can be contacted at email:
[email protected].

Idowu Dauda Oladipo received his B.Sc. (Edu.) degree in Computer Science from
Ekiti State University, Ado-Ekiti, Ekiti State, Nigeria in 2005. He earned his M .Sc. and Ph.D.
degree in Computer Science from University of Ilorin, Ilorin, Nigeria, in 2010 and 2018
respectively. Since 2019, he has been a Lecturer II with the Department of Computer Science,
University of Ilorin, Ilorin, Nigeria. He is the author of more than 20 articles, and more than 10
conference Proceedings. His research interests include Software Engineering, Bioinformatics,
Information Security, Artificial Intelligence, Cyber Security and Computer Education. He is a
member of International Computer Professional Registration Council of Nigeria (M CPN) and
Nigeria Computer Society (M NCS). He can be contacted at email: [email protected],
[email protected].

S ekinat Olaide Adekola received her B.Sc in Computer Science from University of
Ilorin, Ilorin. She received her ordinary National Diploma, OND, in computer science from Osun
State Polytechnic, Iree, Nigeria. She enrolled from the National Institute of Information
Technology, NIIT, in Ajah, Lagos State, Nigeria, with a M icrosoft certification. She can be
contacted at email: [email protected].

Int J Artif Intell, Vol. 11, No. 4, December 2022: 1535-1544