https://doi.org/10.31449/inf.v46i3.4016                                                                                      Informatica 46 (2022) 393-402     393 
Intelligent Analysis and Processing Technology of Big Data Based on 
Clustering Algorithm 
 
Zheng Zheng
1
, Fukai Cao
1*
, Song Gao
2
, Amit Sharma
3
 
1
Jitang College, North China University of Science and Technology, Tangshan,063210, China 
2
Tangshan Power Supply Company, State Grid Jibei Electric Power Co., Ltd, Tangshan,063000, China 
3
Southern Federal University, Russia 
Emails: zhengzheng873@163.com, fukaicao@126.com, songgao56@163.com, amit.amitsharma90@gmail.com  
 
Keywords: Clustering algorithm; Big data intelligence; Smart meter; Project cost; Genetic algorithm  
 
Received: February 15, 2022 
 
An attribute category clustering method based on hierarchical clustering is proposed in order to study 
the big data intelligent analysis and processing technology. The proposed model combines the attribute 
categories with similar fault type distribution, reduces the data dimension, and binarizes it. To address 
the problem of more missing values of continuous data, a data completion method based on attribute 
distribution function is adopted. Through the perspective of selection and estimation of project unit 
price in construction enterprises, this paper summarizes the data mining process facing the 
characteristics of project cost data, and puts forward the method of analyzing and processing project 
cost data based on clustering algorithm. Finally, the processed data sets are subjected to bottom-up 
hierarchical clustering analysis, and finally the ideal analysis results can be obtained. The experimental 
results show that the preprocessing method based on attribute clustering proposed in this paper can 
effectively merge attributes, reduce the dimension after binary transformation and effectively reduce the 
amount of data under the condition of ensuring data information. 
 
Povzetek: S hierarhičnim gručenjem je narejena inteligentna analiza velikih podatkov. 
 
 
1 Introduction  
The hidden value of big data promotes the 
derivation of big data mining technology and methods. 
Big data mining is to mine valuable knowledge for data 
processing through massive multiple data sources. 
Therefore, how to quickly and accurately mine valuable 
knowledge through big data has attracted much attention. 
In fact, data mining is also a decision support process. Its 
common methods mainly include classification, 
clustering, prediction, regression analysis, association 
rules and so on. Clustering is the most key technology. 
Big data is unstructured data, which is difficult and large 
in processing and analysis, making the structural analysis 
mode too complicated, and the traditional data analysis 
cannot effectively process, mine and analyze as shown in 
Figure 1 [1]. The classical methods of cluster analysis 
can be summarized as: partition method, hierarchical 
method, density-based method, grid-based method, 
model-based method, neural network method based on 
computational intelligence, evolutionary computing 
method, fuzzy method and so on, as well as the semi 
supervised clustering method which has attracted much 
attention at present. Recently, the new cluster integration 
method has rapidly become a new research hotspot of 
cluster analysis. The purpose of clustering integration is 
to fuse the results from multiple clustering algorithms to 
obtain higher quality and robust clustering results. The 
method based on graph theory is one of the fast-
developing methods recently. It is a method to realize 
clustering by using the principles of graph theory and 
graphics. Compared with traditional algorithms, this 
algorithm can deal with more complex cluster structures, 
such as nonconvex structures, and can converge to the 
global optimal solution [2]. 
In recent years, with the rapid development of 
network information technology, the era of big data has 
come and penetrated into many fields. There are more 
and more big data application research for specific 
professional fields. However, for the field of project cost, 
this aspect has always been a blank. Every day, with the 
help of the Internet and various project cost systems, a 
large number of project cost data are generated, but there 
is no scientific and accurate processing method to 
process it, so that it is lost in vain. The acquisition and 
transmission of project cost information still rely on the 
traditional way, and the timeliness and accuracy cannot 
meet the needs of today's project management field [3]. 
To process and mine these huge project cost information 
data and provide basis and reference for the decision-
making of project management process, it is not enough 
to rely on manual processing technology. We should 
innovate and apply data mining technology to make full 
use of the value of massive project cost data, so as to 
promote the rapid and healthy development of the 
industry. 
 

394     Informatica 46 (2022) 393-402                                                                                                                              Z. Zheng et al. 
 
Figure 1: Big data intelligent analysis and processing technology 
  
The rest of manuscript is organized as the most 
recent work done is discussed in Section 2. The research 
methodology, optimization of clustering algorithm, 
complexity and project acquisition is presented in 
Section 3. Results and analysis of the proposed model is 
discussed in Section 4 which is followed by the 
conclusion in Section 5.  
 
2 Related work 
In this section various state-of-the-art work in the 
field of big data processing based on clustering algorithm 
is presented.  
  Zhu et al. proposed an initial clustering center 
selection method based on point density, and processed 
outliers specially [4]. Ser et al. proposed an improved 
algorithm to determine the optimal cluster number k by 
calculating the contour coefficient of each object in the 
cluster under different K values, and determine the initial 
cluster center by hierarchical aggregation method [5]. 
Wu proposed a clustering method based on patent 
technology efficacy matrix. This method uses K-means 
to cluster by calculating the similarity of technology, and 
achieves good results. K-medoids and PAM algorithms 
are very effective for small data sets, but they do not 
have good scalability for large data sets [6]. Duan and  
 
 
Wang proposed a new heuristic search algorithm clarans 
algorithm based on PAM [7]. The algorithm finds the  
center point of the representative cluster by random 
search of the graph. Clarans algorithm is the first 
clustering algorithm successfully applied in the field of 
spatial data mining. It overcomes the shortcomings that 
other classical clustering algorithms cannot deal with 
large-scale data sets, but it still fails to solve the problem 
of low execution efficiency. Its time complexity is 2O 
(KN). In order to speed up the execution speed of the 
algorithm, the parallel clarans algorithm based on PVM 
mechanism proposed by Xing and Li effectively 
improves the speed of the algorithm [8]. In the artificial 
neural network, Cai applied the classical hierarchical 
clustering algorithm and partition algorithm to cluster 
SOM, which aims to reduce the computational 
complexity of the classical clustering method [9]. In 
addition, in terms of network application: Xu et al. 
proposed a three-dimensional facial expression clustering 
method based on network, which overcomes the 
shortcomings of limited information contained in data 
and sharp decline in recognition performance in the case 
of two-dimensional facial expression recognition [10]. In 
terms of project cost, Li et al. others established the 
power grid cost management method system and the 
construction framework of cost analysis information 
platform under the big data environment [11]. Shi and 

Intelligent Analysis and Processing Technology of Big…                                                          Informatica 46 (2022) 393-402     395 
Zhu designed the cost management system of mine 
engineering construction project based on cost data [12]. 
Wendong et al. put forward the statistics and analysis 
method of project cost information data under the 
background of big data, and constructed the statistical 
calculation model of project cost information data [13]. 
The evolution of artificial intelligence and Internet of 
Things is considered for several industrial applications 
and contributing towards social life [14-17]. 
 
3 Research methods 
This section includes the project design process, 
structural seismic analysis and detailed modeling steps of 
proposed design.  
As unstructured data, big data is difficult to be 
characterized by two-dimensional logic table of database. 
The multi-dimensional de clustering analysis algorithm 
shows the hidden structure of observation variables 
through the Bayesian network model structure, and 
constructs the logical correlation between leaf nodes and 
other nodes. In this model, multiple hidden variables are 
allowed to exist, corresponding to the corresponding data 
clustering methods. Based on the probability dependence 
between random variables, the multi-dimensional de 
reunion class analysis algorithm analyzes unstructured 
data, and quantitatively describes the reasonable 
distribution with the conditional concept as the carrier. 
The specific flow of data processing is as follows: 
Data preprocessing, that is, data cleaning, avoiding 
noise and solving the problem of data loss. During data 
processing, discrete the continuous values in the attribute 
and convert the data. The data result set and test training 
set are studied, and the data set is divided into two parts: 
data result set and test training set. The classifier is 
constructed by classification algorithm. Through the test 
set, the accuracy evaluation mode is selected to evaluate 
the classifier. The classifier that meets the accuracy 
standard is applied in practice, otherwise it will be 
modified.  
 
 
Figure 2: Data processing flow of cluster analysis model 
 
Word segmentation and document vectorization 
processing, reorganize the continuous word sequence 
according to the established norms to form the word 
sequence. In order to transform the document after word 
segmentation into a pattern that can be recognized and 
processed by computer, it is necessary to quantify the 
word features as the feature vector, which is currently 
processed by vector space model. Feature selection and 
multi-dimensional cluster analysis, word features will 
lead to a certain sparsity and high dimensionality in the 
document vector feature space, so an effective feature 
selection method is selected to reduce the dimension of 
the feature space and further improve the classification 
efficiency and accuracy [18]. The detailed data 
processing steps of the analysis model are shown in 
Figure 2. 
Clearly build a functional model for the 
classification process of big data and unstructured data. 
The problem can be described as a given data set and 
category set which is evaluated using Equation 1 and 2. 
 
 
m
F F F F F F , , , , ,
4 3 2 1
   (1) 
 
m
G G G G , , , , G
3 2 1
   
(2) 
The classification problem is to clarify the function 
mapping to make the data items of the data set map to the 
corresponding categories. Given the big data variable set, 
the variable takes the parent node set as the carrier, the 
carrier correlation between nodes can be characterized by 
a directed graph, that is, for each variable, it can be 
characterized as a node, and each node guides a directed 
edge from each directed node of the parent node set to 
enter the variable. Suppose that the variables of Bayesian 
network are a and b respectively, and X is the node set 
without a and b. once z separates a and b, the conditions 
remain independent based on a given z. The so-called 
isolation and conditional independence show the close 
relationship between the graph theory side and 
probability theory side of Bayesian network. Set to 
classify objects based on the evidence provided by the 
feature vector, then: 
 
j i
x
vi
e
x
vj
e 













 (3) 
    vi e
vi
x
e vj e
vj
x
e















 
(4) 
Decision rules are likelihood test rules which are 
evaluated using Equation 3 and 4. Bayesian network 
reasoning, through probability decomposition, reduces 
the reasoning complexity to localize the operation. 
Through the edge processing and analysis of the 
elimination process, the decision rules can be tested by 
likelihood rate for all given large data sets to obtain the 
minimum error probability calculation samples [19]. 
 

396     Informatica 46 (2022) 393-402                                                                                                                              Z. Zheng et al. 
3.1 Optimization of clustering algorithm  
Based on the function model, an optimized 
clustering algorithm is constructed to divide the overall 
big data into multiple data intervals, which are stored 
through multiple files, and each file represents the 
corresponding interval. After scanning and comparing all 
the data, divide them into multiple sections, and sort and 
remove multiple files. The data quantity of each file is 
1𝑀 and 2𝑀 respectively. After the data is de duplicated, 
cluster analysis is carried out and Bayesian formula is 
used to calculate, that is calculated in Equation 5. 
 
 
2
log , log
M
n f
F
N
e Max
F
N
BIC  





 





  (5) 
𝑀𝑎𝑥 𝛼 log 𝑒 (
𝑁 𝐹 , 𝛼 )represents the effect of data and 
model integration; 𝑓 ( 𝑛 ) 𝑙𝑜𝑔 represents that when the data 
is closely integrated with the model, it should be taken as 
the negative amount of difference, while when combined 
with sparse, it should be taken as the compensation 
amount. Based on the specific specification of Bayesian 
formula and the organic combination of model and data, 
on the basis of meeting the clustering characteristics, it is 
necessary to calculate and analyze the model through 
multi-dimensional clustering algorithm. The input of this 
algorithm contains m objects. The objects in the same 
cluster have high similarity, on the contrary, the 
similarity is small. The algorithm description process is 
shown in Figure 3. 
 
 
Figure 3: Algorithm description process 
3.2 Complexity 
The space cost generated by the new algorithm 
needs to fully consider the characteristic samples of big 
data. If hierarchical clustering is used to optimize the 
clustering algorithm, all clusters to be clustered need to 
be reasonably set according to the serial mode, the total 
clustering time ( 𝑅 ) and the cost ( 𝑛 ) . Then the space 
complexity (W) is expressed in Equation 6. 
 
  m m R n W log
2
  (6) 
In terms of optimization rules, when the model and 
data fusion are sparse, set x and y as the dimension of the 
data set. When dividing attributes, only scan the data set 
at one time, in which 𝑧 identifies clustering data, and the 
results will not be affected by factors such as 
multidimensional space and input order [20, 21]. Then 
the multi-dimensional spatial clustering can be found in 
time through the evaluation of weight and threshold, and 
the amount of calculation can be simplified. The total 
clustering time 𝑢 × 𝑛 is divided by the linear 
arrangement of the consumption time ( 𝑛 ) and the de 
duplication time ( 𝑚 ) ; Total weight removal time 𝑢 × 𝑛 ; 
Time complexity 𝑢 × 𝑚 2 log m, then the total time 
complexity of the algorithm is calculated by Equation 7. 
 
  m m u m u m u m R log
2
      (7) 
3.3 Acquisition of project cost data 
There are two ways to obtain project cost data based 
on big data.  
i. There are generally two methods of internal 
collection in the platform. First method is to 
build a unified project cost information data 
collection template, collect and import the 
relevant data in the platform according to the 
user-defined unified specifications, so as to 
directly convert the target cost data information 
and store it in the local database for backup. The 
second method is to set up fields conforming to 
certain specifications on the relevant cost 
information platform, collect the information of 
the same field and store it in the local database 
[22].  
ii. The specific methods and principles are as 
follows: create a unified data exchange format 
through the corresponding platform interface, 
and realize the information exchange of relevant 
businesses inside and outside the platform. 
According to the collection method and the form of 
price change trend, we generally use the box method to 
process the project cost data studied in this paper. Before 
processing, we must first solve the problem of detection. 
For the detection of noise data, the change of cost data is 
mainly based on the overall change of market economy 
[23]. From the perspective of time series, it changes 
continuously, and is largely affected by the overall 
economic development. Generally, there will be no major 

Intelligent Analysis and Processing Technology of Big…                                                          Informatica 46 (2022) 393-402     397 
fluctuations and changes. We set the percentage of the 
annual change threshold range of cost data to 19%. 
Within the sampling range, the data points exceeding 
20% of the average value are regarded as noise, the 
regression curve is calculated, and its value is re solved 
and corrected [24-26]. Handling method of inconsistent 
data format: to deal with the problem of inconsistent data 
format, the common method is to establish a general data 
acquisition template and collect according to the general 
data template to ensure the consistency of data 
acquisition format. According to the requirements and 
characteristics of data analysis in this paper, the data 
acquisition template is established, as shown in Table 1 
and Table 2. 
 
 
 
Listing Type Accuracy Format Explain 
Region text -- -- Area code 
Number double 1 XXX 
Sample 
number 
Company text -- -- 
Collection 
unit 
Unit Price double 0.02 XX.XX Unit Price 
Single time Date s -- 
Acquisition 
time 
Source Date … -- Data sources 
Table 2: Template description - labor unit price expense 
template 
 
As the material cost accounts for a large proportion 
of the project cost, usually about 0% ~ 70%, the material 
price has a great impact on the specific final settlement 
results and decisions [27]. Therefore, this paper selects 
the material price as the research object, and focuses on 
the specific application of material price data in the fields 
of relevant project cost index prediction, project price 
information analysis and investment estimation. Due to 
the dynamic, massive, multi-source and heterogeneous 
characteristics of project cost big data, we choose K-
means clustering algorithm for specific solution [28]. 
 
4 Results and Analysis 
This section illustrates the analysis of results 
obtained by comparing the seismic forces and finally 
presents its discussion and summary. In this proposed 
model, cluster analyze the quotations of 20 local 
suppliers for composite Portland cement. 
 
Number Region Specifications 
Unit 
Price 
Source 
1 SSX PC32.1 452 
Merchant 
A 
2 SSX PC32.1 326 
Merchant 
A 
3 SSX PC32.1 419 
Merchant 
A 
4 SSX PC32.1 385 
Merchant 
A 
5 SSX PC32.1 453 
Merchant 
A 
6 SSX PC32.1 376 
Merchant 
A 
7 SSX PC32.1 413 
Merchant 
A 
8 SSX PC32.1 306 
Merchant 
A 
9 SSX PC32.1 378 
Merchant 
A 
10 SSX PC32.1 403 
Merchant 
A 
11 SSX PC32.1 487 
Merchant 
A 
… SSX PC32.1 … 
Merchant 
A 
20 SSX PC32.1 346 
Merchant 
A 
Table 3: Data acquisition results 
 
The 20 data listed in Table 3 are combined 
according to the price and serial number to obtain the 
initial data set 𝐴 , 𝐴 is { 𝑥 1
, 𝑥 2
, 𝑥 3
, … , 𝑥 20
}. Before 
calculation, it should be noted that K-Means algorithm 
must give K value before solution, which directly 
Number Region Unit Time Source 
1 Jiangsu yuan January Data survey 
2 Shanghai yuan February Data survey 
3 Beijing yuan March Data survey 
Table 1: Data collection template - labor unit price 
expense template 

398     Informatica 46 (2022) 393-402                                                                                                                              Z. Zheng et al. 
determines the accuracy and efficiency of the algorithm. 
This paper determines the 𝐾 value according to the 
following methods: firstly, compare the distance between 
each sample in the sample data set, select the point 
furthest from other points as the initial center point of the 
calculation according to the calculation results, and then 
determine the value of K through the newly generated 
classification [29, 30]. 
i. Select the two data with the smallest distance in 
the data sequence. In this example, the distance 
between the two points 𝑥 9
 and 𝑥 12
 is the largest. 
Take these two points as the center of the cluster 
for cluster calculation to obtain two cluster sets. 
 
They are: 𝑆 21
= {𝑥 9
, 𝑥 2
, 𝑥 4
, 𝑥 8
, 𝑥 10
, 𝑥 13
, 𝑥 14
, 𝑥 18
} and    
𝑆 22
= { 𝑥 12
, 𝑥 3
, 𝑥 5
, 𝑥 7
, 𝑥 11
, 𝑥 12
, 𝑥 16
, 𝑥 17
}. 
 
ii. Combined with the above clustering calculation 
results, for the two clustering sets, first solve the 
first type of data and cluster center 𝑥 9
 
respectively, for example, to obtain the farthest 
distance of 83, the second type of data and 
cluster center 𝑥 12
 respectively, with the 
maximum distance of 85, and then select the 
point 𝑥 11
 with the maximum distance as the 
third cluster point. 
iii. Recalculate, select 𝑥 9
, 𝑥 12
 and 𝑥 11
 as three 
cluster centers, and calculate three cluster sets as 
follows: 
 
𝑆 31
= {𝑥 9
, 𝑥 2
, 𝑥 10
, 𝑥 20
}, 𝑆 32
= { 𝑥 12
, 𝑥 1
, 𝑥 5
} and 
𝑆 33
= {𝑥 11
, 𝑥 3
, 𝑥 4
, 𝑥 6
, 𝑥 7
, 𝑥 8
, 𝑥 10
, 𝑥 13
, 𝑥 14
, 𝑥 15
} 
 
iv. Calculate the distance between the data elements 
in the three set classes and each cluster center, 
continue the cluster analysis, and then obtain 
four cluster sets [31]. 
v. Based on the above calculation results, the 
cluster numbers of different cluster centers are 
listed, as shown in Table 4. 
 
Serial 
number 
Center point 
Numerical 
value 
Number of 
clusters 
1 
9
x 315 5 
2 
11
x 406 4 
3 
12
x 475 4 
4 
18
x 413 9 
Table 4: Cluster analysis results 
 
According to the analysis of the results of the 
clustering algorithm in Table 4 and Figure 4, point 𝑥 18
 is 
the center with the largest number of clustering samples 
in all clustering centers, so it can better reflect the real 
price of the market compared with other centers [32]. 
Taking this as an example, in the practical application of 
project cost budget and final accounts, we can analyze 
the market price of materials through the data mining 
algorithm proposed in this paper. By analyzing the 
solution results, we can assist relevant personnel to 
accurately grasp the market price information and help 
auditors judge the authenticity of price information in 
time. 
 
 
Figure 4: Results of clustering algorithm 
 
 
Figure 5(a): Result for different size of datasets for 
information loss 
 

Intelligent Analysis and Processing Technology of Big…                                                          Informatica 46 (2022) 393-402     399 
 
Figure 5(b): Result for different size of datasets for 
execute time 
 
Different number of records are separated from the 
grown-up dataset and assess the exhibition of further 
developed anonymity model on various size of datasets, 
as depicted in Figure 5 (a and b). As shown in this figure, 
execute time increment and information loss with the 
increasing size of datasets. Execute time rises quickly, 
yet the incensement of information loss reportedly slows 
progressively. Clearly, the rising size of datasets 
genuinely affects execution time on the grounds that the 
grouping system of finding proportionality classes is 
perplexing and time consuming. 
 
 
Figure 6: Performance comparison of time measured for 
dataset 1 
 
 
Figure 7: Performance comparison of time measured for 
dataset 2 
 
The performance of the proposed clustering scheme 
is measured on two different datasets, dataset 1 i.e., BoW 
(Bag of words) dataset and dataset 2 i.e., HOUSE 
(household electric power consumption) dataset. To 
analyze the performance of clustering cost of proposed 
algorithm we have compared it with existing baseline 
models. The value of 𝑘 is considered as 40 and 80 for 
BoW and HOUSE datasets. Figure 6 and 7 illustrates the 
experimental analysis of HOUSE and Bag of words 
(BoW) datasets and the total running time of proposed 
model is observed. It is observed from the 
experimentation that the proposed model achieves higher 
performance in comparison with K-means ++, K-means 
and K-means || when implanted to execute in parallel.   
 
5 Conclusions 
Different data analysis and mining methods are 
required for different purposes of project cost data 
mining under the background of big data. From the 
perspective of the selection and estimation of engineering 
unit price in construction enterprises, this paper 
summarizes the data mining process facing the 
characteristics of engineering cost data, and puts forward 
the method of analyzing and processing engineering cost 
data based on clustering algorithm. The proposed model 
provides a meaningful exploration for the research of 
massive engineering cost data mining. From the 
experimentation it is analyzed that the proposed 
clustering model achieves better time measurement when 
compared with existing baseline models. The clustering 
models based on computational intelligence are 
proposed. However, these intelligent technologies are not 
organically integrated. Machine learning and data mining 
technology have made great breakthroughs in today's 
academic and industrial circles. Therefore, how to 
integrate various intelligent technologies to give full play 
0
10
20
30
40
50
60
70
80
K-means K-means || K-means ++ Proposed
clustering
method
Time for dataset 1 (mins)
Running time of dataset 1 
Performance analysis for dataset 1
k=40 k=80
0
10
20
30
40
50
60
70
80
K-means K-means || K-means ++ Proposed
clustering
method
Time for dataset 2 (mins)
Running time of dataset 2 
Performance analysis for dataset 2
k=40 k=80

400     Informatica 46 (2022) 393-402                                                                                                                              Z. Zheng et al. 
to the functional characteristics of this kind of algorithm 
applied to cluster analysis is also one of the future 
research directions. 
 
References  
[1] Li, W., & Huang, Q. (2017). Research on intelligent 
avoidance method of shipwreck based on bigdata 
analysis. Polish Maritime Research.   
10.1515/pomr-2017-0125 
[2] Li, L., Wang, J., & Li, X. (2020). Efficiency 
analysis of machine learning intelligent investment 
based on K-means algorithm. Ieee Access, 8, 
147463-147470.  
10.1109/ACCESS.2020.3011366 
[3] Dong-rui, L. (2017). Cluster analysis algorithm 
based on key data integration for cloud 
computing. International Journal of Reasoning-
based Intelligent Systems, 9(3-4), 123-129.  
10.1504/IJRIS.2017.090041 
[4] Zhu, K., Joshi, S., Wang, Q. G., & Hsi, J. F. Y . 
(2019). Guest editorial special section on big data 
analytics in intelligent manufacturing. IEEE 
Transactions on Industrial Informatics, 15(4), 
2382-2385.  
10.1109/TII.2019.2900726 
[5] Del Ser, J., Sanchez-Medina, J. J., & Vlahogianni, 
E. I. (2019). Introduction to the special issue on 
online learning for big-data driven transportation 
and mobility. IEEE Transactions on Intelligent 
Transportation Systems, 20(12), 4621-4623.  
10.1109/TITS.2019.2955548 
[6] Wu, C. (2019, June). Research on Clustering 
Algorithm Based on Big Data Background. 
In Journal of Physics: Conference Series (V ol. 
1237, No. 2, p. 022131). IOP Publishing.  
10.1088/1742-6596/1237/2/022131 
[7] Duan, S., & Wang, Z. (2021). Research on the 
service mode of the university library based on data 
mining. Scientific Programming, 2021.  
https://doi.org/10.1155/2021/5564326 
[8] Xing, Z., & Li, G. (2019). Intelligent classification 
method of remote sensing image based on big data 
in spark environment. International Journal of 
Wireless Information Networks, 26(3), 183-192.  
https://doi.org/10.1007/s10776-019-00440-z 
[9] Cai, Z. M. (2020). Network community partition 
based on intelligent clustering 
algorithm. Компьютерная оптика, 44(6), 985-
989.  
10.18287/2412-6179-CO-724 
[10] Xu, Z., Shi, D., & Tu, Z. (2021). Research on 
diagnostic information of smart medical care based 
on big data. Journal of Healthcare 
Engineering, 2021.  
https://doi.org/10.1155/2021/9977358 
[11] Li, W., Luo, Y ., Tang, C., Zhang, K., & Ma, X. 
(2021). Boosted Fuzzy Granular Regression 
Trees. Mathematical Problems in 
Engineering, 2021.  
https://doi.org/10.1155/2021/9958427 
[12] Shi, F., & Zhu, L. (2019). Analysis of trip 
generation rates in residential commuting based on 
mobile phone signaling data. Journal of Transport 
and Land Use, 12(1), 201-220.  
http://dx.doi.org/10.5198/jtlu.2019.1431 
[13] Wendong, X., Yuanfeng, L., & Deli, C. (2017). 
Algorithm of key data ensemble clustering and 
approximate analysis in cloud 
computing. International Journal of Reasoning-
based Intelligent Systems, 9(3-4), 177-184.  
10.1504/IJRIS.2017.090038 
[14] Singh, P. K., & Sharma, A. (2022). An intelligent 
WSN-UA V-based IoT framework for precision 
agriculture application. Computers and Electrical 
Engineering, 100, 107912. 
https://doi.org/10.1016/j.compeleceng.2022.107912 
[15] Zeng, H., Dhiman, G., Sharma, A., Sharma, A., & 
Tselykh, A. (2021). An IoT and Blockchain‐based 
approach for the smart water management system in 
agriculture. Expert Systems, e12892. 
https://doi.org/10.1111/exsy.12892 
[16] Sharma, A., & Singh, P. K. (2021). UA V‐based 
framework for effective data analysis of forest fire 
detection using 5G networks: An effective approach 
towards smart cities solutions. International 
Journal of Communication Systems, e4826. 
https://doi.org/10.1002/dac.4826 
[17] Sharma, A., Singh, P. K., & Kumar, Y . (2020). An 
integrated fire detection system using IoT and 
image processing technique for smart 
cities. Sustainable Cities and Society, 61, 102332. 
  https://doi.org/10.1016/j.scs.2020.102332 
[18] Tseng, F. H., Cho, H. H., & Wu, H. T. (2019). 
Applying big data for intelligent agriculture-based 
crop selection analysis. IEEE Access, 7, 116965-
116974.  
10.1109/ACCESS.2019.2935564 
[19] Zhao, Y ., Ding, F., Li, J., Guo, L., & Qi, W. (2019). 
The intelligent obstacle sensing and recognizing 
method based on D–S evidence theory for 
UGV . Future Generation Computer Systems, 97, 
21-29.  
https://doi.org/10.1016/j.future.2019.02.003 
[20] Yuan, W., Deng, P., Taleb, T., Wan, J., & Bi, C. 
(2015). An unlicensed taxi identification model 
based on big data analysis. IEEE Transactions on 
Intelligent Transportation Systems, 17(6), 1703-
1713.  
10.1109/TITS.2015.2498180 

Intelligent Analysis and Processing Technology of Big…                                                          Informatica 46 (2022) 393-402     401 
[21] Wang, L. (2021, December). Intelligent analysis of 
accounting information processing under the 
background of big data. In 2021 2nd International 
Conference on Big Data Economy and Information 
Management (BDEIM) (pp. 461-464). IEEE.  
10.1109/BDEIM55082.2021.00100 
[22] Ma, X., Wang, Z., Zhou, S., Wen, H., & Zhang, Y . 
(2018, June). Intelligent healthcare systems assisted 
by data analytics and mobile computing. In 2018 
14th International Wireless Communications & 
Mobile Computing Conference (IWCMC) (pp. 
1317-1322). IEEE.  
10.1109/IWCMC.2018.8450377 
[23] Hu, H., Tang, B., Gong, X., Wei, W., & Wang, H. 
(2017). Intelligent fault diagnosis of the high-speed 
train with big data based on deep neural 
networks. IEEE Transactions on Industrial 
Informatics, 13(4), 2106-2116.  
10.1109/TII.2017.2683528 
[24] Vedavathi, N., Dharmaiah, Ghuram, Venkatadri, 
Kothuru and Gaffar, Shaik Abdul. Numerical study 
of radiative non-Darcy nanofluid flow over a 
stretching sheet with a convective Nield conditions 
and energy activation. Nonlinear Engineering, 
10(1), 159-176, 2021.  
https://doi.org/10.1515/nleng-2021-0012 
[25] Hayat, Tasawar, Ullah, Inayat, Muhammad, 
Khursheed and Alsaedi, Ahmed. Gyrotactic 
microorganism and bio-convection during flow of 
Prandtl-Eyring nanomaterial. Nonlinear 
Engineering, 10(1), 201-212, 2021.  
https://doi.org/10.1515/nleng-2021-0015 
[26] Li, Zhenfang, Gao, Dong, Wu, Chuanji, Lv, 
Guoqing, Liu, Xin, Zhai, Haoran and Huang, 
Zhanfang. Mechanical performance of aerated 
concrete and its bonding performance with glass 
fiber grille. Nonlinear Engineering, 10(1), 240-244, 
2021.   
https://doi.org/10.1515/nleng-2021-0018 
[27] Liang, H., Yun, C., Kan, M. J., & Gao, J. (2019). 
Research and application of element logging 
intelligent identification model based on data 
mining. IEEE Access, 7, 94415-94423.  
10.1109/ACCESS.2019.2928001  
[28] He, Z., He, Y ., Liu, F., & Zhao, Y . (2019). Big data-
oriented product infant failure intelligent root cause 
identification using associated tree and fuzzy 
DEA. IEEE Access, 7, 34687-34698.  
10.1109/ACCESS.2019.2904759 
[29] He, X., Wang, K., Lu, H., Xu, W., & Guo, S. 
(2020). Edge qoe: Intelligent big data caching via 
deep reinforcement learning. IEEE Network, 34(4), 
8-13.  
10.1109/MNET.011.1900393 
[30] Lei, Y ., Jia, F., Lin, J., Xing, S., & Ding, S. X. 
(2016). An intelligent fault diagnosis method using 
unsupervised feature learning towards mechanical 
big data. IEEE Transactions on Industrial 
Electronics, 63(5), 3137-3147. 
10.1109/TIE.2016.2519325  
[31] Srivani, B., Sandhya, N., & Padmaja Rani, B. 
(2020). Literature review and analysis on big data 
stream classification techniques. International 
Journal of Knowledge-Based and Intelligent 
Engineering Systems, 24(3), 205-215.  
10.3233/KES-200042 
[32] Liu, X., Sun, Q., Lu, W., Wu, C., & Ding, H. 
(2020). Big-data-based intelligent spectrum sensing 
for heterogeneous spectrum communications in 
5G. IEEE Wireless Communications, 27(5), 67-73.  
10.1109/MWC.001.1900493 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

402     Informatica 46 (2022) 393-402                                                                                                                              Z. Zheng et al.