Ingénierie des Systèmes d’Information

Vol. 29, No. 3, June, 2024, pp. 1209-1218

Journal homepage: http://iieta.org/journals/isi

Accuracy Rate of Least Square Support Vector Machine Method and Its Various
Modifications: A Forecasting Evaluation on Multi-Type Data
Mahsup1* , Putri Ayu Febriani1, Syaharuddin1 , Vera Mandailina1, Abdillah1 , Ibrahim2
1
Mathematics Education, Universitas Muhammadiyah Mataram, Mataram 83127, Indonesia
2
Geography Education, Universitas Muhammadiyah Mataram, Mataram 83127, Indonesia

Corresponding Author Email: [email protected]

Copyright: ©2024 The authors. This article is published by IIETA and is licensed under the CC BY 4.0 license
(http://creativecommons.org/licenses/by/4.0/).

https://doi.org/10.18280/isi.290337 ABSTRACT

Received: 14 December 2023 This research aims to compare the accuracy levels of the Least Square Support Vector
Revised: 5 May 2024 Machine (LS-SVM) method and its modification with other algorithms in predicting various
Accepted: 16 May 2024 types of data. A quantitative approach with meta-analysis was employed, and the data were
Available online: 20 June 2024 analyzed using JASP software, focusing on Mean Absolute Percentage Error (MAPE),
Effect Size (ES) values, and Summary Effect (SE). The data analysis concludes that,
overall, the LS method exhibits an accuracy rate of 92.7%, categorized as high, with an
Keywords: estimated coefficient value of 0.073. Based on the algorithm used, the analysis results with
accuracy level, algorithm modifications, the LS method achieved an accuracy rate of 87.5%. The LS-SVM method demonstrates a
forecasting, the least squares method higher accuracy level, reaching 95.4%, while the LS-Combination method attains the
highest accuracy rate, namely 95.6%. In data classification, the analysis results indicate the
highest accuracy level in economic and trade data, amounting to 95.6%. For social and
demographic data, the coefficient value is 0.122 with an accuracy rate of 87.8%. Finally, in
agricultural and mining data, the generated accuracy rate is 86.6%. These findings provide
valuable insights into the performance of the LS method and its modifications with other
algorithms in the context of forecasting various types of data.

1. INTRODUCTION residuals.
Research on the LS method has been applied across various
Forecasting is a crucial aspect of conducting predictive domains for economic data forecasting, such as business cycle
analysis [1]. According to Taylor & Letham [2], forecasting is data [11], carbon price [12-15], and stock price estimation [16].
a common task in the field of data science that aids Additionally, studies employing the LS method are prevalent
organizations in capacity planning, goal setting, and anomaly in predicting shallow landslides induced by rainfall [17],
detection. Moran et al. [3] assert that forecasting involves the streamflow forecasting [18, 19], and groundwater surface
ability to predict future occurrences based on the analysis of elevation forecasting [20]. Su et al. [21] utilized the Least
past and current data. On the other hand, prediction is a Square Regression Boosting (LSBoost) method in their
broader concept, referring to assigning a probability research on natural gas price forecasting, obtaining low MSE
distribution to an outcome based on model estimates, and RMSE values of 1.12% and 1.06%, respectively.
applicable to both realized and unrealized outcomes [4]. Furthermore, findings from a study on predicting cow dung
The Least Squares (LS) method is commonly employed for pyrolysis for biochar production indicated that the LS-SVM
estimating coefficients in regression models [5]. According to model outperformed the ANN model with R2 values of 0.96
Fujii [6], the LS method is one statistical approach used to and 0.80, respectively [22].
estimate correlations among various data sets. Meanwhile, Furthermore, the application of the LS method is
Sulaimon Mutiu [7] asserts that the LS method serves as a widespread in climate data analysis, including wind energy
standard approach in regression analysis for solving over- data [23, 24] and air humidity data [25]. Utilizing LS in
determined system approximation, wherein there are more forecasting hourly fluctuations in the Air Quality Index (AQI)
equations than unknowns. The LS method proves valuable in achieved an accuracy level of 93.24%, signifying a remarkably
analyzing experimental data to summarize the effects of high precision [26]. In their study on monthly temperature
factors and test linear contrasts among predictions [8]. estimation using the Partial Least Square (PLS) method, Ertaç
Furthermore, according to Kong et al. [9], the LS method is a et al. [27] reported an RMSE value of 1.80% and a high
way to find the best-fitting function by minimizing the total accuracy level of 94%. Moreover, the LS method finds
sum of squared errors between measured points and the application in forecasting endeavors in the United States and
corresponding straight line. Consistent with Abazid et al. [10], Europe, including the USA [28-32], the Netherlands [33, 34],
the LS method finds widespread application in data fitting, Germany [35, 36], and Brazil [37]. In Peña-Guzmán et al. [38]
aiming for the best fit that minimizes the sum of squared research on forecasting water demand in residential,

1209
commercial, and industrial areas using LS-SVM, an accuracy understanding of the effectiveness of the LS-SVM method and
level of 98% and a percentage error below 12% were achieved, its modifications in forecasting various types of data.
indicating a high level of precision. Consequently, this study not only makes a practical
The LS method yields superior results and presents contribution to selecting the optimal algorithm for time series
reasonable long-term elasticity concerning natural resource forecasting but also strengthens the theoretical foundation in
demand for energy consumption forecasts [39]. The LS the development of more advanced forecasting techniques.
method is extensively employed in forecasting in Asia, The anticipated in outcome of this research is an improvement
including China [40-44], Turkey [45], and Korea [46, 47]. in prediction accuracy, which in turn is expected to yield
Additionally, the integration of the LS method with other significant benefits in decision-making processes.
approaches in the forecasting domain is prevalent. Examples
include LS-SVM for river flow prediction [48], Multi-Task
Learning and LS-SVM for electricity load forecasting [49], LS 2. METHOD
Support Vector Machine Coupled with Data-Preprocessing
Techniques (LSSVM-DWA) for river flow forecasting, This research adopts a quantitative approach utilizing meta-
Volterra-LS Support Vector Machine Model Combined with analysis to delve into the error rates of the LS, LS-SVM, and
Signal Decomposition (EMD-LMD-LSSVM-Volterra) for the combination of LS with other algorithms in various
hourly solar radiation forecasting [50], Fuzzy Clustering and forecasting domains such as economics, industry, social
Least Square Support Vector Machine Optimized with Wolf sciences, demography, agriculture, and mining. Typically, the
Pack Algorithm (FC-WPA-LSSVM) for short-term load classification function of LS-SVM can be formulated as
forecasting at electric bus charging stations [51], and Least follows [55]:
Square Support Vector Machine, Deep Belief Network,
Singular Spectrum Analysis, and Locality-Sensitive Hashing 𝑓(𝑥) = 𝑆𝑖𝑔𝑛 (𝑊 𝑇 𝜑(𝑥) + 𝑏) (1)
(LSSVM-DBN-SSA-LSH) for wind power estimation [52].
Research outcomes by Tien Bui et al. [53] indicate that the where, 𝜑(x) is referred to as a nonlinear function mapping
predictive strength of LSSVM-BC surpasses that of SVM, from the input space X to a high-dimensional feature space.
achieving an accuracy rate of 93.8% in predicting landslides The coefficients w and b are obtained by minimizing the upper
due to rainfall. The IGA-LS-SVM algorithm utilized by Lin et bound of the generalization error. Thus, Eq. (1) can be derived
al. [54] provides a more significant MSE value of 0.83%, by solving the following optimization problem:
making it suitable for short-term power load prediction.
Numerous studies on the application of the LS method, 1 1
𝑙

either standalone or in combination with various algorithms, min 𝑤 𝑇 𝑤 + 𝛾 ∑ 𝜉𝑙2

2 2 (2)
have yielded diverse accuracy levels. Earlier research findings 𝑖=1
𝑇
indicate that the LS method, combined with algorithms such s. t. 𝑦𝑖 = 𝑤 𝜑(𝑥𝑖 ) + 𝑏 + 𝜉𝑖 , (𝑖 = 1,2, … , 𝑙)
as Support Vector Machine (SVM), fuzzy logic, Gravitational
Search Algorithm (GSA), Genetic Programming (GP), where, 𝜉𝑖 is the error variable and 𝛾 is the penalty parameter.
Evolutionary Seasonal Decomposition (ESD), and Signal By employing the Lagrangian function and Karush-Kuhn-
Decomposition (SD), achieves high accuracy levels based on Tucker (KKT) conditions for optimality in Eq. (2), we can
Mean Absolute Percentage Error (MAPE) values. However, to derive the final classification solution of the primal problem as
date, there has been a lack of focus on accuracy levels based follows:
on MAPE values. The combination of the LS method with
other techniques also impacts the number of iterations, training 𝑙

duration, and testing of data. Therefore, the author has 𝑓(𝑥) = 𝑆𝑖𝑔𝑛 (∑ 𝑤𝑖 𝐾(𝑥, 𝑥𝑖 ) + 𝑏) (3)
undertaken data collection to discuss the research outcomes of 𝑖=1
the LS-SVM method, both in its standalone form and in
combination with other algorithms. The objective of this In Eq. (3), K represents the kernel function, which serves to
research is to assess the comparative accuracy levels of the LS simplify the utilization of mapping. The research methodology,
method and its modifications across various types of illustrated in Figure 1, outlines the procedural steps undertaken
forecasted data. This research aims to provide a deeper in this study.

Figure 1. Research procedure

1210
 The data collection process involved searching for relevant Based on the analysis of Figure 2 and Figure 3, the data
research findings in both national and international indexed analysis results reveal the utilization of three classification
databases. Inclusion criteria encompassed studies conducted intervals for data forecasting: LS method, LS-SVM method,
between 2013-2023, employing keywords such as “Prediction, and LS modification with other algorithms. The LS method
Forecasting, LS, LS-SVM”, and featuring statistical data demonstrates commendable performance with 19 data points,
including data volume and accuracy parameters (MAD, MSE, exhibiting an average Mean Absolute Percentage Error
MAPE, or RMSE). Conversely, exclusion criteria targeted (MAPE) of 0.13, an average Effect Size (ES) of 0.064, and an
quantitative studies and articles available only in full text. All average Standard Error (SE) of 0.035, categorizing it as highly
relevant articles were downloaded and stored in a designated proficient.
folder. In the second phase, each article was reviewed to
ascertain (1) the input data size (N) during prediction, training,
or testing stages, and (2) the value of Mean Absolute
Percentage Error (MAPE). Subsequently, coding and
tabulation were conducted, encompassing (1) publication year;
(2) author names; (3) type of forecasted data; (4) country; (5)
forecasting method (LS); (6) data size (N); and (7) Mean
Absolute Percentage Error (MAPE) values. The accuracy
criteria for MAPE, as outlined by Yadav and Nath [56], are
presented in Table 1.
Table 1 presents criteria indicating the accuracy level based
on Mean Absolute Percentage Error (MAPE) values to classify
whether forecast outputs can be accepted or rejected. MAPE
values <10% indicate highly accurate forecasting, 10% <
MAPE <20% denotes good forecasting, 20% < MAPE <50%
suggests acceptable forecasting, and MAPE >50% indicates
inaccurate forecasting. Subsequently, Mean Absolute
Percentage Error (MAPE) values were transformed into effect Figure 2. Data quantity based on method and data
size (ES), and finally, ES values were converted into summary classification
effect (SE) using the following formula:

Table 1. Criteria for MAPE value

Percentage Decimal Category

<10% <0.1 Very accurate
10%≤MAPE<20% 0.1≤MAPE<0.2 Accurate
20%≤MAPE<50% 0.2≤MAPE<0.5 Worthy
≥50% ≥0.5 Not accurate

1 + 𝑀𝐴𝑃𝐸
𝐸𝑆 = 0.5 × 𝐿𝑁 ( ) (4)
1 − 𝑀𝐴𝑃𝐸

𝐸𝑆(1 − 𝐸𝑆)
𝑆𝐸 = √ (5)
𝑁
Figure 3. Average values of MAPE, ES, and SE based on
Next, the data is tabulated in Microsoft Excel and saved in method classification
CVS (Macintosh) format. The stored data is then uploaded to
the JASP software for coefficient determination, rank p-value In the LS-SVM method, based on dataset of 10 entries, the
testing, and forest plot generation. The output from JASP is average MAPE stands at 0.057, with an average ES of 0.104
interpreted to assess the accuracy level based on the MAPE and an average SE of 0.034, also falling within the excellent
values pf the Least Square method in forecasting, both overall category. Meanwhile, the LS method combine with other
and when employing conventional and combined methods. algorithms showcases outstanding performance, yielding an
Finally, conclusions are drawn based on the conducted average MAPE of 0.053, an ES of 0.084, and an SE of 0.035,
analysis. leveraging a dataset of 25 entries. Furthermore, the data is
categorized into three intervals: economic and industrial (38
data points), social and demographic (9 data points), and
3. RESULT AND DISCUSSION agricultural and mining (7 data points). The MAPE values for
each data point are delineated in Figure 4.
3.1 The search result In Figure 4, the analysis results indicate that the average
Mean Absolute Percentage Error (MAPE) value reaches 0.081,
A total of 140 data points were collected for the study, with signifying a remarkably high level of accuracy in forecasting.
54 meeting the specified criteria. There were 19 articles on the Furthermore, the data reveals that the minimum recorded
LS method, 10 articles on the LS-SVM method, and 25 articles MAPE value is 0.001, while the maximum value reaches 0.412.
on the LS method combined with other algorithms. The This information serves as an indicator of the model's
distribution of data is illustrated in Figure 2 and Figure 3. reliability, considering that lower MAPE values correspond to

1211
more accurate forecasts. The subsequent step involves model), coefficients, p-value rank test values, and funnel plot
determining the values of effect size (ES) and standard error to ascertain the average error rate based on the predetermined
(SE) based on MAPE values and the number of data points (N), model. The resulting JASP output can be found in Tables 2-4.
as illustrated in Figure 5.
Table 2. Fixed and random effect
3.2 Homogeneity test and publication bias test
Q df p
Based on the ES and SE values of each dataset, the data Omnibus test of model coefficients 35.816 1 <.001
were processed using JASP software to determine the category Test of residual heterogeneity 1290.756 53 <.001
of the method employed (Fixed Effect or Random Effect

Figure 4. Summary of MAPE values

Figure 5. Values of ES and SE for each data point

1212
 Table 3. Coefficients

95% Confidence Interval

Estimate Standard Error z p Lower Upper
Intercept 0.073 0.012 5.985 <.001 0.049 0.097

Table 4. p-value of rank correlation bias. In other words, the results of the p-Rank test < α = 0.05
suggest no indication of bias, as seen in Table 4. The forest
Rank Correlation Test for Funnel Plot Asymmetry plot value is 0.07 [0.05, 0.10], indicating an error rate of 7.3%,
Kendall's τ p thus demonstrating an overall accuracy level of 92.7% for the
Rank test 0.222 0.019 LS method. Subsequently, the author divides the data based on
the classification of methods and algorithms used. The JASP
According to Table 2, the obtained information indicates outputs based on these classifications are presented in Table 5.
that the input data fall into the heterogeneous category, with a
residual heterogeneity test value of 1,290.756 and a significant
variation among the data, indicating the complexity in the
characteristics of the dataset used. Furthermore, from Table 3
an estimation coefficient value of 0.073 with a significant p-
value < 0.001 is obtained for the LS method. This illustrates
that the LS method demonstrates a low level of error in
predicting or forecasting data, affirming the reliability of this
method in predictive analysis. Meanwhile, in Table 4 the p-
value of rank correlation is found to be 0.019 > 0.001,
indicating no indication of publication bias within the dataset
utilized. Thus, the data from the 54 articles included in the
study are deemed sufficiently representative to draw
conclusions from this research, without significant influence
from publication bias.
In Figure 6, the JASP software output reveals that the funnel
plot of the overall data indicates no research bias, marked by
all circles enclosed or declared as not indicating publication Figure 6. Funnel plot of overall data

Table 5. MAPE category

Method N Estimate Category of MAPE Z p Forest Plot Accuracy Level

LS 19 0.125 12.5% Accurate 4.478 <.001 0.13 [0.07, 0.18] 87.5%
LS-SVM 10 0.046 4.6% High Accurate 2.191 0.028 0.05 [0.00, 0.09] 95.4%
LS-Combination 25 0.044 4.4% High Accurate 4.455 <.001 0.04 [0.02, 0.06] 95.6%

From the results in Table 5, it is evident that the error rate signifying that the research sample meets minimal standards.
for the LS method with N = 19 is 13% (0.07-0.18), and the In other words, the p-Rank results for each algorithm are
estimated value is 0.125 with a p-value < 0.001, signifying an greater than 0.001 (p-value > 0.001), implying no indication
accuracy level of 87.5%. Furthermore, the error rate for LS- of publication bias. Here, the author divides the data based on
SVM with N = 10 yields a result of 4.6% (0.00-0.09), an the publication year, data quantity, and data classification in
estimated value of 0.046, and a p-value of 0.028 < 0.05, each study.
indicating that LS-SVM has an accuracy level of 95.4%. This
aligns with the research by Sun and Liang [57], where short-
term electricity load forecasting using LS-SVM achieved a
low MAPE value of 2.34%. However, in the research
conducted by Biswas et al. [58], experimental result indicate
that the performance of the MARS model slightly outperforms
that of the LS-SVM model. For LS-Combination with N = 25,
the error rate is 4.4% (0.02-0.06), the estimated value is 0.044,
and the p-value is < 0.001, implying that LS-Combination also
has an accuracy level of 95.6% in predicting data. Consistent
with the findings of Niu and Dai [59], the proposed short-term
load forecasting model with EMD-GRA-MPSO-LS-SVM
outperforms other models with a MAPE value of 1.1%. In
contrast to the study by Yang et al. [60] utilizing the PLSE-
RVM method, where the accuracy rate reached 63.3%. The
distribution pattern can be observed in Figure 7.
From the JASP output, it is evident that the plot trajectories
of the three algorithms indicate no indication of publication
bias, as marked by the absence of open circles, meaning no
missing studies. Additionally, all circles are enclosed, (a) Funnel plot of LS method

1213
 power output prediction using the ESD-LS-SVR method yield
a MAPE value of 0.078 or 7.8%, indicating a very accurate.
Furthermore, the data quantity is divided into two categories:
equal to or less than 120 (N≤120) and greater than 120
(N>120). For N≤120, the coefficient value is 0.062, the p-
Rank test value is < 0.001, and the forest plot is 0.06 [0.03,
0.09], signifying an error rate of 6.2%. This error rate is higher
than the findings of Tian’s [63] study, which reported a MAPE
value of 3.79%. In the case of N>120, the coefficient value is
0.096, the p-Rank test value is 0.007, and the forest plot is 0.10
[0.03, 0.16], with a minimum value of 3% and a maximum
value of 16%. These results are higher than the findings of
Yang et al. [64] with a MAPE value of 0.0206 or 2.06% in the
worthy category. Electricity load forecasting results with a
data quantity of 84 and a MAPE value of 0.006 or 0.6%, as
well as a MAPE value of 0.008 or 0.8%, both fall into the very
accurate category, according to the studies [65, 66],
respectively.
In the realm of data classification, three distinct categories
(b) Funnel plot of LS-SVM method emerge: economic and trade data, social and demographic data,
and agricultural and mining data. Focusing on economic and
trade data, where N = 38, the coefficient stands at 0.044, the
p-Rank at 0.056, and the forest plot at 0.04 [0.03, 0.06]. With
a minimum value of 3% and a maximum of 6%, this attests to
an accuracy level reaching 95.6%. This aligns with Mustaffa
et al. [67] research on commodity price forecasting, boasting
a MAPE value of 5.5%, and stock prediction with a MAPE of
0.8% [68]. Tang et al. [69] estimated building material prices
with an MSE of 2.44% and MAPE of 2.11%.
Transitioning to social and demographic data, where N = 9,
a notable error rate is evident with a coefficient of 0.122, p-
Rank of 0.477, and a forest plot of 0.12 [0.05, 0.20], signifying
an accuracy level of 87.8%. In Song et al. [70] study on water
quality prediction using LS-SVM, a MAPE of 5.2%
underscores highly accurate forecasts. Chia et al. [71]
achieved a MAPE of 5.6% in water quality index prediction
through the SMWOA-LSSVM method. Turning to
agricultural and mining data with N = 7, a coefficient of 0.134,
p-Rank of 0.562, and forest plot of 0.13 [0.02, 0.24] showcase
(c) Funnel plot of LS-combination with other algorithms an accuracy level of 86.6%. Demonstrating very accurate,
Yang et al. [72] forecasted short-term electricity load with AS-
Figure 7. Funnel plot of each method GCLSSVM, yielding a MAPE of 0.006 or 0.6%.
In this study, the results demonstrate that the Least Square
3.3 Moderator variable test Support Vector Machine (LS-SVM) method and its
modifications exhibit a significant level of accuracy in
This study delves into the performance of LS method and forecasting various types of data. The LS-SVM method proves
its modifications in predicting various types of data. By to achieve a high level of accuracy reaching 95.4% in this
categorizing the data based on publication year, data quantity, research. This finding aligns with previous studies
and classification. The data research offers a comprehensive highlighting the superiority of the LS-SVM method in short-
understanding of how the method’s performance varies across term electricity load forecasting. However, it is noteworthy
different contexts. Additionally, supplementary analysis is that the study also finds that the MARS model slightly
conducted by considering moderator variables, as illustrated in outperforms LS-SVM in certain contexts.
Table 6. Furthermore, the findings of this research provide valuable
In the publication year interval, it is divided into two periods: insights for practitioners and researchers in understanding the
2013-2018 and 2019-2023. For the years 2013-2018, the applicability and limitations of the forecasting methods
coefficient value is 0.044, the p-Rank test value is 0.044, and employed. This study highlights potential avenues for future
the forest plot is 0.04 [0.03, 0.06], indicating a very accurate research, including refining algorithmic modifications, further
with an accuracy level of 95.6%. For the years 2019-2023, the investigating error rates in specific data categories, and
coefficient value obtained is 0.099, the p-Rank test value is developing more sophisticated forecasting models. It is
0.149, and the forest plot is 0.10 [0.06, 0.14], meaning the important to acknowledge the limitations of this study, such as
accuracy level reaches 90.1% in predicting data. This aligns sample size and the scope of data types considered which
with the research by Yan and Chowdhury [61] with a MAPE could be addressed in future research to strengthen findings
value of 0.11 or 11%. The findings of Lin and Pai [62] on solar and broaden our understanding of forecasting methods.

1214
 Table 6. Moderator variable

Variables Interval N Coefficient p-Rank Test Forest Plot Category of MAPE

2013-2018 29 0.044 0.044 0.04 [0.03, 0.06] Very Accurate
Publication year
2019-2023 25 0.099 0.149 0.10 [0.06, 0.14] Very Accurate
≤120 38 0.062 < 0.001 0.06 [0.03, 0.09] Very Accurate
Data quantity
> 120 16 0.069 0.007 0.10 [0.03, 0.16] Very Accurate
Economic and trade data 38 0.044 0.056 0.04 [0.03, 0.06] Very Accurate
Data classification Social and demographic data 9 0.122 0.477 0.12 [0.05, 0.20] Accurate
Agricultural and mining data 7 0.134 0.562 0.13 [0.02, 0.24] Accurate

4. CONCLUSIONS [3] Moran, K.R., Fairchild, G., Generous, N., Hickmann, K.,
Osthus, D., Priedhorsky, R., Hyman, J., Valle, S.Y.D.
The utilization of the LS method, combined with (2016). Epidemic forecasting is messier than weather
algorithmic combinations, has proven instrumental in forecasting: The role of human behavior and internet data
optimizing regression models for accurate data forecasting. streams in epidemic forecast. The Journal of Infectious
This approach, integrating various methodologies, Diseases, 214(s4): S404-S408.
significantly enhances prediction precision. The analysis https://doi.org/10.1093/infdis/jiw375
reveals an overall accuracy rate of 92.7% for the LS method, [4] Hegre, H., Metternich, N.W., Nygård, H.M.,
categorized as high, with an estimated coefficient of 0.073. Wucherpfennig, J. (2017). Introduction: Forecasting in
Notably, algorithm-specific examination demonstrates an peace research. Journal of Peace Research, 54(2): 113-
accuracy rate of 87.5% and an estimation value of 0.125 for 124. https://doi.org/10.1177/0022343317691330
the LS method. Conversely, the LS-SVM method exhibits a [5] Guo, Y., Nazarian, E., Ko, J., Rajurkar, K. (2014).
lower accuracy rate of 4.6% (0.00-0.09) with an estimation Hourly cooling load forecasting using time-indexed
value of 0.046, indicative of low error rates. ARX models with two-stage weighted least squares
In the case of LS-Combination, an error value of 4.4% is regression. Energy Conversion and Management, 80: 46-
obtained, ranging from 2% to 6%, with an estimation value of 53. https://doi.org/10.1016/j.enconman.2013.12.060
0.044, resulting in an impressive accuracy level of 95.6%. [6] Fujii, K. (2018). Least squares method from the view
Further accuracy analysis within the data range (N≤120) yields point of deep learning. Advances in Pure Mathematics,
a coefficient of 0.062 and a forest plot of 0.06 [0.03, 0.09], 8(5): 485-493. https://doi.org/10.4236/apm.2018.85027
signifying an accuracy level of 93.8%. For datasets exceeding [7] Sulaimon Mutiu, O. (2015). Application of weighted
120 (N>120), a coefficient of 0.096, a p-Rank value of 0.007, least squares regression in forecasting. International
and a forest plot of 0.10 [0.03, 0.16] signify an accuracy level Journal of Recent Research in Interdisciplinary Sciences,
of 90.4%, categorized as high. 2(3): 45-54.
Moreover, the data classification analysis underscores the [8] Lenth, R.V. (2016). Least-squares means: The R package
highest accuracy in economic and trade data at 95.6%, lsmeans. Journal of Statistical Software, 69(1): 1-33.
accompanied by a coefficient of 0.044, a p-Rank value of https://doi.org/10.18637/jss.v069.i01
0.056, and a forest plot of 0.04 [0.03, 0.06]. Conversely, social [9] Kong, M., Li, D., Zhang, D. (2019). Research on the
and demographic data exhibit a considerable error, with a application of improved least square method in linear
coefficient of 0.122, p-Rank value of 0.477, and an accuracy fitting. IOP Conference Series: Earth and Environmental
level of 78%. Meanwhile, agricultural and mining data show a Science, 252(5): 052158. https://doi.org/10.1088/1755-
coefficient of 0.134, a p-Rank value of 0.562, and a forest plot 1315/252/5/052158
of 0.13 [0.02, 0.24], indicating an accuracy level of 86.6%. [10] Abazid, M., Abdulrahman, A., Samine, S. (2018). Least
The findings underscore the efficacy of the LS method in squares methods to forecast sales for company.
enhancing prediction accuracy across various datasets, International Journal of Scientific & Engineering
particularly in economic and trade contexts. The study Research, 9(6): 864-868.
highlights the practical implications of these findings in [11] Ferrara, L., van Dijk, D. (2014). Forecasting the business
improving forecasting techniques, with potential future cycle. International Journal of Forecasting, 30(3): 517-
research directions focusing on refining algorithmic 519. https://doi.org/10.1016/j.ijforecast.2013.12.001
combinations and addressing error rates in specific data [12] Zhu, B., Ye, S., Wang, P., Chevallier, J., Wei, Y.M.
categories, such as social and demographic datasets. (2022). Forecasting carbon price using a multi-objective
least squares support vector machine with mixture
kernels. Journal of Forecasting, 41(1): 100-117.
REFERENCES https://doi.org/10.1002/for.2784
[13] Jianwei, E., Ye, J., He, L., Jin, H. (2021). A denoising
[1] Ahmad, A.S., Hassan, M.Y., Abdullah, M.P., Rahman, carbon price forecasting method based on the integration
H.A., Hussin, F., Abdullah, H., Saidur, R. (2014). A of kernel independent component analysis and least
review on applications of ANN and SVM for building squares support vector regression. Neurocomputing, 434:
electrical energy consumption forecasting. Renewable 67-79. https://doi.org/10.1016/j.neucom.2020.12.086
and Sustainable Energy Reviews, 33: 102-109. [14] Zhu, B., Han, D., Wang, P., Wu, Z., Zhang, T., Wei, Y.M.
https://doi.org/10.1016/j.rser.2014.01.069 (2017). Forecasting carbon price using empirical mode
[2] Taylor, S.J., Letham, B. (2017). Forecasting at scale. decomposition and evolutionary least squares support
PeerJ Preprints, 5: e3190v2. vector regression. Applied Energy, 191: 521-530.
https://doi.org/10.7287/peerj.preprints.3190v2 https://doi.org/10.1016/j.apenergy.2017.01.076

1215
[15] Zhu, B., Wei, Y. (2013). Carbon price forecasting with a 253-265. https://doi.org/10.1007/s00704-014-1235-7
novel hybrid ARIMA and least squares support vector [28] Ahn, S.C., Bae, J. (2022). Forecasting with Partial Least
machines methodology. Omega, 41(3): 517-524 Squares when a Large Number of Predictors are
https://doi.org/10.1016/j.omega.2012.06.005 Available. SSRN. https://doi.org/10.2139/ssrn.4248450
[16] Rahimi, Z.H., Khashei, M. (2018). A least squares-based [29] Chin, W., Cheah, J.H., Liu, Y., Ting, H., Lim, X.J., Cham,
parallel hybridization of statistical and intelligent models T.H. (2020). Demystifying the role of causal-predictive
for time series forecasting. Computers & Industrial modeling using partial least squares structural equation
Engineering, 118: 44-53. modeling in information systems research. Industrial
https://doi.org/10.1016/j.cie.2018.02.023 Management & Data Systems, 120(12): 2161-2209.
[17] Pham, B.T., Tien Bui, D., Dholakia, M.B., Prakash, I., https://doi.org/10.1108/IMDS-10-2019-0529
Pham, H.V. (2016). A comparative study of least square [30] Kim, J.M., Jung, H. (2018). Time series forecasting using
support vector machines and multiclass alternating functional partial least square regression with stochastic
decision trees for spatial prediction of rainfall-induced volatility, GARCH, and exponential smoothing. Journal
landslides in a tropical cyclones area. Geotechnical and of Forecasting, 37(3): 269-280.
Geological Engineering, 34: 1807-1824. https://doi.org/10.1002/for.2498
https://doi.org/10.1007/s10706-016-9990-0 [31] Satija, A., Caers, J. (2015). Direct forecasting of
[18] Adnan, R.M., Yuan, X., Kisi, O., Adnan, M., Mehmood, subsurface flow response from non-linear dynamic data
A. (2018). Stream flow forecasting of poorly gauged by linear least-squares in canonical functional principal
mountainous watershed by least square support vector component space. Advances in Water Resources, 77: 69-
machine, fuzzy genetic algorithm and M5 model tree 81. https://doi.org/10.1016/j.advwatres.2015.01.002
using climatic data from nearby station. Water Resources [32] F.Hair Jr, J., Sarstedt, M., Hopkins, L., G.Kuppelwieser,
Management, 32: 4469-4486. V. (2014). Partial least squares structural equation
https://doi.org/10.1007/s11269-018-2033-2 modeling (PLS-SEM) An emerging tool in business
[19] Londhe, S.N., Gavraskar, S. (2018). Stream flow research. European Business Review, 26(2): 106-121.
forecasting using least square support vector regression. https://doi.org/10.1108/EBR-10-2013-0128
Journal of Soft Computing in Civil Engineering, 2(2): [33] Dijkstra, T.K., Henseler, J. (2015). Consistent partial
56-88. https://doi.org/10.22115/SCCE.2017.96717.1024 least squares path modeling. MIS Quarterly, 39(2): 297-
[20] Moravej, M., Amani, P., Hosseini-Moghari, S.M. (2020). 316. https://doi.org/10.25300/MISQ/2015/39.2.02
Groundwater level simulation and forecasting using [34] Henseler, J., Sarstedt, M. (2013). Goodness-of-fit indices
interior search algorithm-least square support vector for partial least squares path modeling. Computational
regression (ISA-LSSVR). Groundwater for Sustainable Statistics, 28: 565-580. https://doi.org/10.1007/s00180-
Development, 11: 100447. 012-0317-1
https://doi.org/10.1016/j.gsd.2020.100447 [35] Ringle, C.M., Sarstedt, M., Mitchell, R., Gudergan, S.P.
[21] Su, M., Zhang, Z., Zhu, Y., Zha, D. (2019). Data-driven (2020). Partial least squares structural equation modeling
natural gas spot price forecasting with least squares in HRM research. The International Journal of Human
regression boosting algorithm. Energies, 12(6): 1094. Resource Management, 31(12): 1617-1643.
https://doi.org/10.3390/en12061094 https://doi.org/10.1080/09585192.2017.1416655
[22] Cao, H., Xin, Y., Yuan, Q. (2016). Prediction of biochar [36] Homburg, C., Klarmann, M., Vomberg, A. (2022).
yield from cattle manure pyrolysis via least squares Handbook of Market Research. Springer.
support vector machine intelligent approach. https://doi.org/10.1007/978-3-319-05542-8
Bioresource Technology, 202: 158-164. [37] Nanni, M.R., Cezar, E., Silva Junior, C.A.D., Silva,
https://doi.org/10.1016/j.biortech.2015.12.024 G.F.C., da Silva Gualberto, A.A. (2018). Partial least
[23] Ding, M., Zhou, H., Xie, H., Wu, M., Liu, K.Z., squares regression (PLSR) associated with spectral
Nakanishi, Y., Yokoyama, R. (2021). A time series response to predict soil attributes in transitional
model based on hybrid-kernel least-squares support lithologies. Archives of Agronomy and Soil Science,
vector machine for short-term wind power forecasting. 64(5): 682-695.
ISA Transactions, 108: 58-68. https://doi.org/10.1080/03650340.2017.1373185
https://doi.org/10.1016/j.isatra.2020.09.002 [38] Peña-Guzmán, C., Melgarejo, J., Prats, D. (2016).
[24] Xu, M., Lu, Z., Qiao, Y., Min, Y. (2017). Modelling of Forecasting water demand in residential, commercial,
wind power forecasting errors based on kernel recursive and industrial zones in Bogotá, Colombia, using least-
least-squares method. Journal of Modern Power Systems squares support vector machines. Mathematical
and Clean Energy, 5(5): 735-745. Problems in Engineering, 2016: 5712347.
https://doi.org/10.1007/s40565-016-0259-7 https://doi.org/10.1155/2016/5712347
[25] Yang, Z.C. (2019). Hourly ambient air humidity [39] Costa, O.L., de Oliveira Ribeiro, C., Ho, L.L., Rego, E.E.,
fluctuation evaluation and forecasting based on the least- Parente, V., Toro, J. (2020). A robust least square
squares Fourier-model. Measurement, 133: 112-123. approach for forecasting models: an application to
https://doi.org/10.1016/j.measurement.2018.10.002 Brazil’s natural gas demand. Energy Systems, 11(4):
[26] Yang, Z.C. (2020). DCT-based least-squares predictive 1111-1135. https://doi.org/10.1007/s12667-019-00351-1
model for hourly AQI fluctuation forecasting. Journal of [40] Wang, Y., Hao, X., Wu, C. (2021). Forecasting stock
Environmental Informatics, 36(1). 58-69. returns: A time-dependent weighted least squares
https://doi.org/10.3808/jei.201800402 approach. Journal of Financial Markets, 53: 100568.
[27] Ertaç, M., Firuzan, E., Solum, Ş. (2015). Forecasting https://doi.org/10.1016/j.finmar.2020.100568
Istanbul monthly temperature by multivariate partial [41] Gong, X.L., Liu, X.H., Xiong, X., Zhuang, X.T. (2019).
least square. Theoretical and Applied Climatology, 121: Forecasting stock volatility process using improved least

1216
 square support vector machine approach. Soft squares support vector machines and differential
Computing, 23: 11867-11881. evolution optimization: A case study in Central Vietnam.
https://doi.org/10.1007/s00500-018-03743-0 International Journal of Digital Earth, 9(11): 1077-1097.
[42] Zhao, X., Chen, X., Xu, Y., Xi, D., Zhang, Y., Zheng, X. https://doi.org/10.1080/17538947.2016.1169561
(2017). An EMD-based chaotic least squares support [54] Lin, B.D., Zhang, X.Y., Zhang, M., Li, H., Lu, G.Q., L.
vector machine hybrid model for annual runoff (2021). Improved genetic algorithm-based research on
forecasting. Water, 9(3): 153. optimization of least square support vector machines: an
https://doi.org/10.3390/w9030153 application of load forecasting. Soft Computing, 25(18):
[43] Yang, Z.C. (2016). Discrete cosine transform ‐based 11997-12005. https://doi.org/10.1007/s00500-021-
predictive model extended in the least‐squares sense for 05674-9
hourly load forecasting. IET Generation, Transmission & [55] Wang, S., Shang, W. (2014). Forecasting direction of
Distribution, 10(15): 3930-3939. China security index 300 movement with least squares
https://doi.org/10.1049/iet-gtd.2016.0689 support vector machine. Procedia Computer Science, 31:
[44] Yang, Z.C. (2014). Modeling and forecasting monthly 869-874. https://doi.org/10.1016/j.procs.2014.05.338
movement of annual average solar insolation based on [56] Yadav, V., Nath, S. (2017). Forecasting of PM 10 using
the least-squares Fourier-model. Energy Conversion and autoregressive models and exponential smoothing
Management, 81: 201-210. technique. Asian Journal of Water, Environment and
https://doi.org/10.1016/j.enconman.2014.02.033 Pollution, 14(4): 109-113. https://doi.org/10.3233/AJW-
[45] Kaytez, F. (2020). A hybrid approach based on 170041
autoregressive integrated moving average and least- [57] Sun, W., Liang, Y. (2015). Least-squares support vector
square support vector machine for long-term forecasting machine based on improved imperialist competitive
of net electricity consumption. Energy, 197: 117200. algorithm in a short-term load forecasting model. Journal
https://doi.org/10.1016/j.energy.2020.117200 of Energy Engineering, 141(4): 04014037.
[46] Jabbari, A., Bae, D.H. (2020). Improving ensemble https://doi.org/10.1061/(asce)ey.1943-7897.0000220
forecasting using total least squares and lead-time [58] Biswas, R., Rai, B., Samui, P., Roy, S.S. (2020).
dependent bias correction. Atmosphere, 11(3): 300. Estimating concrete compressive strength using MARS,
https://doi.org/10.3390/atmos11030300 LSSVM and GP. Engineering Journal, 24(2): 41-52.
[47] Jung, H.C., Kim, J.S., Heo, H. (2015). Prediction of https://doi.org/10.4186/ej.2020.24.2.41
building energy consumption using an improved real [59] Niu, D., Dai, S. (2017). A short-term load forecasting
coded genetic algorithm based least squares support model with a modified particle swarm optimization
vector machine approach. Energy and Buildings, 90: 76- algorithm and least squares support vector machine
84. https://doi.org/10.1016/j.enbuild.2014.12.029 based on the denoising method of empirical mode
[48] Adnan, R.M., Liang, Z., Heddam, S., Zounemat- decomposition and grey relational analysis. Energies,
Kermani, M., Kisi, O., Li, B. (2020). Least square 10(3): 408. https://doi.org/10.3390/en10030408
support vector machine and multivariate adaptive [60] Yang, C., Zhang, Y., Huang, M., Liu, H. (2021).
regression splines for streamflow prediction in Adaptive dynamic prediction of effluent quality in
mountainous basin using hydro-meteorological data as wastewater treatment processes using partial least
inputs. Journal of Hydrology, 586: squares embedded with relevance vector machine.
124371.https://doi.org/10.1016/j.jhydrol.2019.124371 Journal of Cleaner Production, 314: 128076.
[49] Tan, Z., De, G., Li, M., Lin, H., Yang, S., Huang, L., Tan, https://doi.org/10.1016/j.jclepro.2021.128076
Q. (2020). Combined electricity-heat-cooling-gas load [61] Yan, X., Chowdhury, N.A. (2014). Mid-term electricity
forecasting model for integrated energy system based on market clearing price forecasting using multiple least
multi-task learning and least square support vector squares support vector machines. IET Generation,
machine. Journal of Cleaner Production, 248: 119252. Transmission & Distribution, 8(9): 1572-1582.
https://doi.org/10.1016/j.jclepro.2019.119252 https://doi.org/10.1049/iet-gtd.2013.0610
[50] Wang, Z., Tian, C., Zhu, Q., Huang, M. (2018). Hourly [62] Lin, K.P., Pai, P.F. (2016). Solar power output
solar radiation forecasting using a volterra-least squares forecasting using evolutionary seasonal decomposition
support vector machine model combined with signal least-square support vector regression. Journal of
decomposition. Energies, 11(1): 68. Cleaner Production, 134: 456-462.
https://doi.org/10.3390/en11010068 https://doi.org/10.1016/j.jclepro.2015.08.099
[51] Zhang, X. (2018). Short-term load forecasting for electric [63] Tian, Z., Li, S. (2017). A network traffic prediction
bus charging stations based on fuzzy clustering and least method based on IFS algorithm optimised LSSVM.
squares support vector machine optimized by wolf pack International Journal of Engineering Systems Modelling
algorithm. Energies, 11(6): 1449. and Simulation, 9(4): 200-213.
https://doi.org/10.3390/en11061449 https://doi.org/10.1504/IJESMS.2017.087553
[52] Zhang, Y., Le, J., Liao, X., Zheng, F., Li, Y. (2019). A [64] Yang, F., Li, M., Huang, A., Li, J. (2014). Forecasting
novel combination forecasting model for wind power time series with genetic programming based on least
integrating least square support vector machine, deep square method. Journal of Systems Science and
belief network, singular spectrum analysis and locality- Complexity, 27(1): 117-129.
sensitive hashing. Energy, 168: 558-572. https://doi.org/10.1007/s11424-014-3295-2
https://doi.org/10.1016/j.energy.2018.11.128 [65] Dongxiao, N., Tiannan, M., Bingyi, L. (2017). Power
[53] Tien Bui, D., Pham, B.T., Nguyen, Q.P., Hoang, N.D. load forecasting by wavelet least squares support vector
(2016). Spatial prediction of rainfall-induced shallow machine with improved fruit fly optimization algorithm.
landslides using hybrid integration approach of Least- Journal of Combinatorial Optimization, 33: 1122-1143.

1217
 https://doi.org/10.1007/s10878-016-0027-7 particle swarm optimization. Architectural Engineering
[66] Daut, M.A.M., Hassan, M.Y., Abdullah, H., Rahman, and Design Management, 15(3): 196-212.
H.A., Abdullah, M.P., Hussin, F. (2017). An improved https://doi.org/10.1080/17452007.2018.1556577
building load forecasting method using a combined least [70] Song, C., Yao, L., Hua, C., Ni, Q. (2021). A water quality
square support vector machine and modified artificial prediction model based on variational mode
bee colony. ELEKTRIKA-Journal of Electrical decomposition and the least squares support vector
Engineering, 16(1): 1-5. machine optimized by the sparrow search algorithm
https://doi.org/10.11113/elektrika.v16n1.22 (VMD-SSA-LSSVM) of the Yangtze River, China.
[67] Mustaffa, Z., Yusof, Y., Kamaruddin, S.S. (2014). Environmental monitoring and assessment, 193(6): 363.
Enhanced artificial bee colony for training least squares https://doi.org/10.1007/s10661-021-09127-6
support vector machines in commodity price forecasting. [71] Chia, S.L., Chia, M.Y., Koo, C.H., Huang, Y.F. (2022).
Journal of Computational Science, 5(2): 196-205. Integration of advanced optimization algorithms into
https://doi.org/10.1016/j.jocs.2013.11.004 least-square support vector machine (LSSVM) for water
[68] Lai, L., Liu, J. (2014). Support vector machine and least quality index prediction. Water Supply, 22(2): 1951-
square support vector machine stock forecasting models. 1963. https://doi.org/10.2166/ws.2021.303
Computer Science and Information Technology, 2(1): [72] Yang, A., Li, W., Yang, X. (2019). Short-term electricity
30-39. https://doi.org/10.13189/csit.2014.020103 load forecasting based on feature selection and Least
[69] Tang, B.Q., Han, J., Guo, G.F., Chen, Y., Zhang, S. Squares Support Vector Machines. Knowledge-Based
(2019). Building material prices forecasting based on Systems, 163: 159-173.
least square support vector machine and improved https://doi.org/10.1016/j.knosys.2018.08.027

1218