https://doi.org/10.31449/inf.v49i11.7302 Informatica 49 (2025) 109–120 109 
Strengthening Accounting Information Systems with Advanced Big 
Data Mining Algorithms: Innovative Exploration of Data Cleaning 
and Conversion Automation 
Shu Yang 
Department of International Trade, Shanxi Vocational College of Tourism, Taiyuan, 030031, Shanxi, China 
Email: yangshu_sxcj@163.com 
Keywords: accounting information system, big data mining algorithms, decision support, financial performance 
Received: October 11, 2024 
With the rapid development of big data technology, accounting information systems are facing 
unprecedented challenges in processing and analyzing massive financial data. In order to improve the 
efficiency and accuracy of data processing, this article deeply explores the application of big data mining 
algorithms in optimizing accounting information systems. By introducing advanced big data mining 
algorithms, accounting information systems have achieved automation in data cleaning, transformation, 
and analysis, significantly reducing manual intervention and improving data processing efficiency. This 
article compares the performance of different big data mining algorithms in the analysis of accounting 
informationization risk transactions. Through practical verification, we found that the selected algorithm 
performs well in terms of accuracy, reaching over 95%, which is a significant improvement compared to 
traditional methods. Meanwhile, in terms of computation time, the algorithm has also demonstrated 
significant advantages, reducing computation time by over 30% when processing datasets of the same 
size. These performance improvements not only improve the operational efficiency of accounting 
information systems, but also provide enterprises with more accurate and timely financial information. In 
addition, this article also conducted a survey on the intelligent management of accounting information 
systems. We collected valuable opinions on the current status of accounting information intelligent system 
management by distributing survey questionnaires to in-service MBA and MPAcc students. The survey 
results show that over 76% of accounting personnel and almost all management personnel (91.18%) agree 
with the intelligent features of accounting information systems and believe that establishing a separate 
accounting knowledge base or knowledge management system is necessary. This discovery further 
emphasizes the importance of optimizing accounting information systems and provides direction for future 
research. 
Povzetek: Analizirani so algoritmi rudarjenja velikih podatkov za izboljšanje računovodskih 
informacijskih sistemov. Raziskava poudarja potrebo po inteligentnih sistemih v računovodstvu ter 
potrjuje pozitivne učinke na finančno odločanje in obvladovanje tveganj v podjetjih. 
 
 
1 Introduction 
The intelligent accounting information system can 
display the financial and operational status of enterprises 
in different regions through technologies such as 
geographic information systems (GIS), providing support 
for the global layout and risk management of enterprises. 
The system can penetrate into various business areas of 
the enterprise, provide targeted financial management and 
decision support, and help the enterprise achieve refined 
management [1]. By utilizing big data analysis and 
artificial intelligence technology, the system can deeply 
mine and analyze the integrated data, discover patterns 
and trends in the data, and provide scientific basis for 
enterprise decision-making. On this basis, researching and 
constructing an intelligent agent dynamic accounting 
information platform has important practical significance 
and future value. In the stage of computerized accounting, 
computers are widely used in daily accounting operations. 
Such as setting accounting subjects, filling out accounting 
vouchers, registering accounting books, cost accounting, 
and preparing accounting statements. This not only greatly 
reduces the workload of accounting personnel, but also 
improves the accuracy and efficiency of accounting [2]. It 
is not like computerized accounting, where manual 
accounting is simply simulated through computer 
technology. At present, the existing accounting 
information systems in China are mainly used to process 
transactions that have already occurred [3]. 
Data warehouse technology is a special data storage 
technology that can extract data from numerous databases 
and convert it into a special new format, providing 
decision analysis for decision-makers. It is a collection of 
data that reflects continuous historical changes. The data 
source of a data warehouse is not unique and often 
includes multiple sources, including internal and external 
data of the enterprise (such as survey reports, documents, 
etc.) [4]. It reorganizes, arranges, and stores a large 
amount of historical or current comprehensive data as 
needed, providing random queries, comprehensive data, 
and trend analysis information over time. In the intelligent 

110 Informatica 49 (2025) 109–120 S. Yang 
interactive visualization accounting information system, 
to establish a big data analysis platform, Hadoop, a big 
data processing architecture, can be considered [5]. 
The emergence of big data mining algorithms 
provides strong technical support for the optimization of 
accounting information systems. These algorithms are 
based on artificial intelligence and machine learning 
technologies, capable of automatically extracting valuable 
information from massive data, discovering correlations 
and patterns between data, and providing decision support 
and risk management for enterprises. Through the 
application of big data mining algorithms, accounting 
information systems can achieve automated and 
intelligent data processing and analysis, greatly improving 
work efficiency and accuracy, reducing labor costs and 
error rates. The application of big data mining algorithms 
in optimizing accounting information systems has broad 
prospects and profound significance [6]. Firstly, through 
automated processing and analysis of data, big data 
mining algorithms can significantly improve the 
efficiency of accounting information systems. Traditional 
accounting information systems require a significant 
amount of manpower and time to process data, while big 
data mining algorithms can automate these tasks, greatly 
shortening the data processing cycle. Secondly, big data 
mining algorithms can improve the accuracy of 
accounting information systems [7]. Traditional data 
processing methods often have errors and omissions, 
while big data mining algorithms can ensure the accuracy 
and reliability of data through precise algorithms and 
models [8]. Therefore, this article designs a big data 
mining algorithm to provide predictive analysis and 
decision support models for accounting information 
systems. By mining and analyzing historical data, 
algorithms can predict future financial trends and market 
changes, provide valuable information and suggestions for 
enterprises, and help them make wiser decisions. 
The contribution points of research innovation are as 
follows: 
1) The innovation proposed in this article lies in the 
successful application of big data mining algorithms to 
accounting information systems, achieving automation 
and intelligence in data processing. 
2) This article uses advanced model recognition 
technology and data integrity verification algorithms to 
verify accounting information data files item by item. This 
method not only effectively identifies anomalies and 
errors in the data, but also ensures the integrity and 
consistency of the data. Meanwhile, by comparing with 
other algorithms, the identity verification technology 
proposed in this article has been significantly improved, 
further enhancing the security of accounting information 
systems. 
In addition to innovations in data processing and 
validation, this article also proposes the application of big 
data mining algorithms in predictive analysis and decision 
support. By mining patterns and trends in data, big data 
mining algorithms can provide valuable predictive 
information and decision-making recommendations for 
enterprises. 
 
2 Related work 
In the field of tax management, the application of 
artificial intelligence technology is gradually 
demonstrating its enormous potential and value. Some 
scholars have explored how artificial intelligence can help 
modernize tax management systems, and through deep 
learning and data mining, artificial intelligence can 
identify potential tax risks. And provide warnings to help 
the tax department take timely prevention and control 
measures [9]. This can not only reduce the occurrence of 
tax violations, but also ensure the security and stability of 
national taxation. In terms of auditing, artificial 
intelligence can quickly identify potential financial fraud 
and violations through data analysis and pattern 
recognition. Artificial intelligence can also automate audit 
procedures, reduce manual intervention, and improve the 
accuracy and efficiency of audits [10]. In terms of 
accounting, artificial intelligence can automate daily 
accounting and reporting work, reducing the workload of 
accounting personnel. Artificial intelligence can also 
provide accurate financial forecasts and decision-making 
recommendations for enterprises based on historical data 
and predictive models [11]. 
The application of financial technology based on 
artificial Internet of Things, especially the development of 
big data management algorithms, has brought 
unprecedented opportunities and challenges to the 
financial industry. It explores the application and 
development of big data management algorithms based on 
artificial Internet of Things in financial technology [12]. 
In the field of financial technology, the application of 
AIoT enables financial institutions to collect and analyze 
large amounts of financial data in real-time, improving the 
efficiency and accuracy of financial services. In the credit 
field, big data management algorithms are used to help 
financial institutions quickly assess borrowers' credit 
status, achieve accurate risk pricing and risk control [13]. 
In the investment field, big data management algorithms 
can provide personalized investment advice and risk 
management solutions for investors by analyzing and 
predicting historical data. The application of artificial 
intelligence technology in the field of accounting is 
becoming increasingly widespread. From automated 
accounting processing, intelligent auditing to predictive 
analysis, artificial intelligence technology is gradually 
changing traditional accounting methods. Automated 
accounting processing is one of the earliest applications of 
artificial intelligence technology in the field of 
accounting. Through machine learning and natural 
language processing techniques, artificial intelligence can 
automatically recognize, classify, and input financial data, 
greatly improving work efficiency and accuracy [14]. 
Intelligent auditing utilizes artificial intelligence 
technology to deeply analyze and mine large amounts of 
financial data, helping auditors quickly identify potential 
financial fraud and violations, and improve audit 
efficiency and accuracy [15]. 
The management accounting information system 
should be a comprehensive system that integrates 
budgeting, performance evaluation, analysis and 

Strengthening Accounting Information Systems with Advanced Big… Informatica 49 (2025) 109–120 111 
forecasting, and decision support. Its core lies in not only 
acquiring and processing traditional financial accounting 
information, but also incorporating more non-financial 
information to meet the diverse decision-making needs of 
management [16]. By deeply integrating artificial 
intelligence and big data technology, this system can 
efficiently extract valuable information from massive data 
and provide accurate and timely decision-making basis for 
management. Empowered by big data and artificial 
intelligence, management accounting information systems 
can achieve intelligent analysis of data, reveal hidden 
patterns and trends behind the data, and help enterprises 
make more informed decisions. At the same time, research 
achievements from companies such as Informatica in data 
management and integration (such as the application of 
platforms like Informatica PowerCenter in data cleaning, 
transformation, and integration). This provides strong 
technical support for the management accounting 
information system, ensuring data quality and 
consistency, and further enhancing the system's decision 
support capabilities [17]. In order to effectively solve a 
series of large and complex data problems currently 
existing in enterprises, and to effectively handle various 
current and historical data distributed inside and outside 
the enterprise, it is necessary to establish various themed 
database management systems. In this way, accounting 
personnel and decision-makers can access the integrated 
database system through various front-end analysis 
software tools. And make various decisions based on 
accurate and comprehensive historical information, 
quickly putting the overall solution of the business plan 
into practice [18]. 
 
Table 1: Key results gap in reference research 
Reference 
research 
Key Results Technological gap The necessity of carrying out this 
work 
[9] Artificial intelligence helps modernize 
tax management systems and identify 
tax risks 
The scalability and accuracy 
of tax risk identification 
models need to be improved 
Develop more efficient and accurate 
tax risk identification algorithms to 
improve tax management efficiency 
[10] Artificial intelligence quickly identifies 
financial fraud and violations, 
automates audit procedures 
The automation level of audit 
procedures is limited, and 
real-time performance needs 
to be improved 
Enhance the intelligence and real-time 
performance of audit procedures, 
reduce manual intervention, and 
improve audit efficiency 
[11] Artificial intelligence automates 
accounting and reporting work, 
providing financial forecasting and 
decision-making recommendations 
The accuracy and real-time 
performance of the prediction 
model need to be further 
optimized 
Develop more accurate predictive 
models to provide valuable decision 
support for enterprises 
[12] The application of AIoT in financial 
technology improves the efficiency and 
accuracy of financial services 
The adaptability of big data 
management algorithms in 
the financial field needs to be 
strengthened 
Exploring big data management 
algorithms applicable to the financial 
sector to enhance the level of financial 
services 
[13] Big data management algorithms are 
used for credit risk assessment, 
achieving risk pricing and risk control 
The real-time and accuracy 
of credit risk assessment 
models need to be improved 
Develop more efficient credit risk 
assessment algorithms to enhance the 
risk management capabilities of 
financial institutions 
[14] Artificial intelligence automatically 
recognizes, classifies, and inputs 
financial data to improve accounting 
efficiency 
The accuracy and efficiency 
of accounting automation 
processing need to be further 
improved 
Optimize accounting automation 
processing flow, improve work 
efficiency and accuracy 
[15] Intelligent auditing utilizes artificial 
intelligence technology to deeply 
analyze financial data and improve 
audit efficiency 
The depth and breadth of 
audit data analysis need to be 
expanded 
Strengthen the ability to analyze audit 
data, improve audit efficiency and 
accuracy 
[16] The management accounting 
information system should obtain 
relevant information beyond financial 
accounting information and provide 
decision support 
The integration and 
intelligence level of 
information systems need to 
be improved 
Building a more intelligent 
management accounting information 
system to provide comprehensive 
decision support 
[17] Blockchain ensures the authenticity and 
integrity of accounting records, 
improves transparency and audit 
efficiency 
The application of 
blockchain technology in the 
field of accounting needs to 
be further expanded and 
optimized 
Explore more application scenarios of 
blockchain technology in the 
accounting field to enhance the 
transparency and audit efficiency of 
accounting work 
[18] Establish a thematic database 
management system, provide an 
integrated database system, and support 
decision-making 
The integration and real-time 
performance of database 
management systems need to 
be strengthened 
Build a more integrated and real-time 
database management system to 
provide comprehensive information 

112 Informatica 49 (2025) 109–120 S. Yang 
support for accounting personnel and 
decision-makers 
 
From the above table, it can be seen that although 
artificial intelligence technology has made significant 
progress in fields such as tax management, auditing, 
accounting, and financial technology, there are still many 
technological gaps, such as shortcomings in scalability, 
accuracy, and real-time performance. It is particularly 
important to carry out this work in order to effectively 
solve these problems. By conducting in-depth research on 
the application of big data mining algorithms in 
accounting information systems, we can further optimize 
algorithms for tax risk identification, automation of audit 
procedures, financial forecasting, and decision 
recommendations. It improves the accuracy and efficiency 
of accounting automation processing, strengthens audit 
data analysis capabilities, builds a more intelligent 
management accounting information system, and explores 
more application scenarios of blockchain technology in 
the accounting field. These tasks will help improve the 
overall performance of accounting information systems 
and provide more comprehensive, accurate, and real-time 
information support for enterprises. 
 
3 System model construction 
3.1. Design of an intelligent accounting analysis 
management system 
At present, accounting informatization in China has 
been around for nearly 40 years, but there has not been a 
significant breakthrough in the field of accounting 
informatization. This article believes that there are two 
main reasons: firstly, the market focuses more on the 
construction of accounting business level systems, while 
neglecting research on accounting information systems 
related to management decision-making; Secondly, due to 
the fact that Chinese enterprises often rely solely on 
experience to make decisions in the risk management 
process, without the support of corresponding data, it 
often leads to major mistakes. It can be inferred that 
establishing a risk management information system and 
incorporating it into the management accounting 
information system is extremely necessary. Based on this, 
this article believes that an intelligent modern 
management accounting information system should 
include four subsystems (Figure 1). 
 
 
Figure 1: Intelligent management accounting information 
system 
 
From its functional perspective, it is equivalent to a 
traditional accounting information system, with the goal 
of improving data processing capabilities. After 
introducing data mining technology, this subsystem can 
effectively enhance data processing capabilities and 
obtain a large amount of accounting information. 
Enterprise managers can use performance management 
information systems to set effective performance goals for 
each employee and connect the enterprise strategy with 
each employee.  
 
 
Figure 2: Different stages of accounting informatization 
 
Figure 2 shows the different stages of accounting 
informatization. Data mining techniques can help analyze 
historical data, determine which indicators have the 
greatest impact on organizational performance, and adjust 
the indicators and weights in the balanced scorecard 
accordingly. By mining patterns in historical data, future 

Strengthening Accounting Information Systems with Advanced Big… Informatica 49 (2025) 109–120 113 
performance trends can be predicted, providing a basis for 
organizations to set reasonable performance goals. Data 
mining technology can monitor organizational 
performance data in real-time, detect abnormal situations 
and issue warnings, helping managers take timely 
measures to make adjustments. Through in-depth mining 
and analysis of performance data, problems and 
deficiencies in the organization's operations and 
management can be identified, providing support for 
formulating improvement measures and making scientific 
decisions. Accounting decision support systems can help 
decision-makers in enterprises better utilize their financial 
information to make effective decisions. It is based on 
modern management science and information technology, 
utilizing techniques such as quantitative economics, 
operations research, and control theory to establish 
relevant models, while utilizing computer technology to 
solve semi-structured and unstructured accounting 
problems. 
3.2. Overall architecture of intelligent accounting 
analysis and management system 
The accounting analysis management system can be 
divided into four subsystems: data extraction, data 
warehouse storage, information processing, and 
information visualization display (Figure3). With the help 
of these four subsystems, the goal of integrating, 
collaborating, sharing, controlling, intelligentizing, and 
integrating accounting business management can be 
achieved. The task of the accounting data warehouse 
storage subsystem is to store accounting data not only in 
the database according to the requirements of accounting 
transaction processing, but also synchronously in a data 
warehouse that is convenient for data mining according to 
the theme; The task of the accounting information 
processing subsystem is to process accounting data into 
accounting information; The task of the accounting 
information visualization display subsystem is to provide 
the processing results to information users in various 
visual ways through human-computer interaction. 
 
Figure 3: Composition of accounting analysis management system 
 
The data is stored in a standardized format in a 
unified Data Warehouse (DW) to achieve effective data 
sharing (Figure 4). The data in DW is theme oriented, such 
as sales, production, or customers. The data is organized 
around a specific theme, and when targeted towards theme 
users, it can determine how the business is conducted and 
its reasons. Based on DW technology, enterprise managers 
can discover the relationships between accounting 
information through OLAP technology. With the help of 
DM technology, they can understand the hidden value 
behind data, discover potential favorable information, and 
provide favorable support for enterprise decision-making. 
 
 
Figure 4: Accounting data retrieval system 
 
 

114 Informatica 49 (2025) 109–120 S. Yang 
3.3. Construction of accounting information 
system model under big data mining 
algorithms 
In this article, we propose and validate an 
optimization scheme for accounting information systems 
based on big data mining algorithms, which has achieved 
significant results in feature selection and model accuracy. 
By comparing the data accuracy and training time before 
and after feature selection, we found that the model after 
feature selection significantly reduced data complexity 
and training time while maintaining high accuracy. This 
discovery emphasizes the importance of feature selection 
in improving model performance and efficiency. 
Compared with classical discrete models and project 
response theory models, our model performs well in 
accuracy prediction and testing results on three open 
datasets. Although classical discrete models may still be 
applicable in some cases, our model demonstrates higher 
accuracy in most cases. This may be due to our model 
adopting more advanced algorithms and a more 
comprehensive feature set, which can better capture 
complex patterns and relationships in the data. It should 
be noted that the accuracy of the theoretical model 
reflected in the project is the lowest, which may be due to 
certain issues in model design or implementation. In order 
to improve the accuracy of the model, it may be necessary 
to re-examine its assumptions, parameter settings, or data 
processing methods in the future to ensure that it better 
adapts to practical application scenarios. In the 
preprocessing step, the first step is to remove duplicate 
data, process missing values, correct erroneous data, etc. 
Convert data of different dimensions to the same 
dimension to improve the training effectiveness of the 
model. Traditional accounting information systems often 
require a significant amount of time and manpower when 
processing large amounts of data. And big data mining 
algorithms can automatically process and analyze data, 
greatly improving data processing efficiency. This enables 
accounting personnel to obtain accurate information faster 
and provide more timely data support for enterprise 
decision-making. Figure 5 shows the data mining process 
for accounting and financial management systems. 
 
 
Figure 5: Data mining process for accounting and financial management systems 
 
In each sample 
m in
S X
i
 in the minority class 
sample 
min
S , we used the Nearest Neighbors class of the 
sklearn library to find the nearest neighbors of a given 
sample. Then, we randomly selected a point from these 
nearest neighbors and calculated the difference in feature 
vectors between the sample and this point. Synthesize a 
minority class sample 
new
X as: 
( )   − + =
j i i new
Y X X X                    (1) 
If it is assumed that the upsampling rate is n , then 
n points should be randomly selected from the k nearest 
neighbor points found, and then this method is used to 
synthesize artificial samples for each minority class 
sample. 
Real time segmentation of time series data has broad 
application prospects in the field of accounting. Through 
reasonable testing and validation methods, we can ensure 
the accuracy of segmentation and the effectiveness of the 
model, thereby providing strong support for financial 
management and decision-making in enterprises. The 

Strengthening Accounting Information Systems with Advanced Big… Informatica 49 (2025) 109–120 115 
application of real-time segmentation of time series data 
in the field of accounting is crucial, as it can help us better 
understand the changing trends of financial data and make 
accurate predictions and analyses based on them. Real 
time segmentation is based on certain statistical feature 
indicators for time-series data. Time series data is a 
collection of data points arranged in chronological order, 
usually used to represent the trend of a certain 
phenomenon over time. In the field of accounting, time 
series data may include daily transaction volume, monthly 
revenue, annual profit, etc. The goal of real-time 
segmentation is to divide time series data into multiple 
data segments, so that the data sequences in each segment 
follow the same statistical model. In this way, we can 
apply corresponding statistical methods to analyze and 
predict each data segment: 
 
( ) ( ) ( )      = , , , , ,
1 c i
t x t x t x X            (2) 
Where 
c
t is the current moment. Data stream 
segmentation is the segmentation of 
X
 into a series of 
consecutive non-empty data segments 
     , , , , ,
1 s j
X X X , where: 
 
( ) ( ) ( )   s j X n t x t x t x X
j j l j j j
, , 2 , 1 , , , , , ,
, 1 ,
 =   
      (3) 
 
j l j l j c i l j
n l t t t t t t , , 2 , 1 , , , , , , ,
1 , , 1 ,
 =    
+
＜ (4) 
 
j
n is the data sequence length of 
j
X . 
s
X is the 
data segment containing the current data ( )
c
t x , which is 
called the current data segment. 
Let the data in 
j
X be described by a linear 
regression model: 
 
( ) ( ) ( )  
j
n j j j j
t t t t t f t x
, 1 ,
, , , ,   + =  
      (5) 
 
The linear regression model corresponding to data 
segment 
j
X is: 
( )
j j j
b t a t f + =  ,                          (6) 
The model parameter vector is: 
 
T
j j j
b a , =                               (7) 
( ) ( )
x
m
i
i
u i H  = 
                            (8) 
( )  
i
v i CHAL , =                             (9) 
Generate a unique serial number for each file block. 
This serial number can be generated based on a timestamp, 
file hash, or other unique identifier. Meanwhile, generate 
a corresponding random number for each file block. This 
random number can be used for subsequent encryption, 
signing, or other security operations. After grouping and 
numbering all user file blocks, merge these file blocks in 
a certain order into one large file or data stream. This 
merging process can be based on the sequence number of 
file blocks or other sorting rules to ensure that the merged 
files are ordered. 

=
+ =
n
i
r i k i k
m v
1
,
                            (10) 
In the formula: 
r
 represents the random number 
generated by the cloud storage server for each user during 
each verification process. Compute the signature: 
 
 
= =








 =
K
k
n
i
k
v
i k
r
i
1 1
,
                           (11) 
 
To further improve the reliability of data, error 
correction techniques can be used, such as solving linear 
system equations for error correction. This method is 
usually applied in data communication and storage, and 
when there may be a small number of errors in the 
received data, linear equations are solved to recover the 
original data. In the context of cloud storage, if errors 
occur during data transmission or storage, linear system 
equations can be used for error correction through the 
following steps: 
 

=
−  = =
v
k
j
k k j
k n j X Y s
1
, , 1 ,             (12) 
The error value 
k
Y can be determined where: 
 
k
i
k
X  =                                 (13) 
 
Accompanied by: 
 
( ) ( ) ( )

=
−  = = = =
v
k
i
j
i
j j
j
k n j e e R S
k
k 1
, , 1 ,         (14) 
 
Corresponding 
k
i

 and 
k
i
e to 
k
X and 
k
Y 
respectively, then: 
 
( ) ( ) ( )

=
−  = = = =
v
k
j
k k
j j
j
k n j X Y e R S
1
, , 1 ,        (15) 
That is, 
k
X gives the wrong position and 
k
Y gives 
the wrong value. 
 
4 Intelligent investigation of 
accounting information system 
management 
In order to study the necessity of accounting 
knowledge management and compare the views of 
management and accounting fields on knowledge 
management, this article specifically selected in-service 
MBA (representing management personnel) and MPAcc 
(representing accounting personnel) students as the 
objects to conduct a survey on the current situation of 
accounting information intelligent system management. 
Hope to obtain the opinions of middle and senior 
management and finance department personnel on 
accounting knowledge management in the enterprise. A 
total of 70 survey questionnaires were distributed, and 55 

116 Informatica 49 (2025) 109–120 S. Yang 
valid questionnaires were collected, with an effectiveness 
rate of 78.57%, including 21 MPAcc questionnaires and 
34 MBA questionnaires. Among the MPAcc survey 
respondents, 57.14% had more than 5 years of work 
experience, and 61.90% in the finance department; 
Among the MBA survey respondents, 55.88% have more 
than 5 years of work experience, and 41.18% are middle-
level managers, which can basically reflect the actual 
situation of the enterprise. The main results are as follows: 
More than 76% of MPAcc and MBA employees 
believe that personal experience accumulation contributes 
to the development of departmental business and the 
company. This indicates that implicit knowledge related 
to personal experience is of high importance to 
departmental business, and there is a clear need for 
intelligent information management. 
 
Table 1: Intelligent attributes of accounting 
information systems 
Intelligent 
attributes of accounting 
information systems 
MPAcc MBA 
Existence 76.19% 91.18% 
Not existence 23.81% 8.82% 
Total 100% 100% 
 
Table 1 addresses the intelligent management issues 
inherent in accounting knowledge compared to other 
knowledge. The survey results show that 76.19% of 
accounting personnel agree with the intelligent 
characteristics of accounting information systems, while 
almost all managers (91.18%) agree with the 
professionalism and specialization of accounting 
knowledge. managers believe that even if a company 
already has an information management system, it is still 
necessary to establish a separate knowledge base or 
knowledge management system that is suitable for the 
specialization of accounting knowledge. This also 
indicates that although knowledge management has been 
promoted for many years, accounting knowledge 
management in the financial field is still necessary. It is 
also not difficult to see that management personnel have a 
higher demand for accounting knowledge management 
than accounting personnel, indicating that accounting 
personnel have a high level of professionalism in 
providing useful decision-making information, and it is 
necessary to make up for it through intelligent accounting 
information management methods. 
 
5 Result analysis  
By selecting features from different training datasets, 
this article ultimately used 70 financial indicators out of 8 
features for data analysis 
 
 
 
 
 
 
 
Table 2: Accuracy before and after feature selection 
 Quantity 
of features 
Accuracy Training 
time 
Data 
before 
feature 
selection 
70 75.91% 0.0469 
Data after 
feature 
selection 
8 82.16% 0.0411 
 
According to the results in Table 2, it can be clearly 
observed that the accuracy of the feature set after feature 
selection has increased by 6.25%, and there are 8 feature 
subsets after feature selection, which is much lower than 
the original feature set. This reduces the complexity of the 
data, improves the learning efficiency of learners, and thus 
reduces training time. The significant reduction from 70 
original financial indicators to 8 feature subsets not only 
simplifies the data, but also reduces the complexity of the 
model. Fewer features mean that the model needs to learn 
and process less information, which usually leads to faster 
training speed and better generalization ability. The 
adjustment parameters of the model were tested for 
motion accuracy prediction on the recommendation 
model, item reflection theory model, and classical discrete 
model in three open datasets. From Figure 6, it can be seen 
that although the results of the classical discrete model are 
not significantly different from the model proposed in this 
paper, this does not mean that the classical model has no 
advantages or value. In practical applications, classical 
models may still have certain applicability, especially in 
specific contexts or datasets. Therefore, when choosing a 
model, it is necessary to weigh specific requirements and 
data characteristics. The theoretical model reflected in the 
project has the lowest accuracy, which may be due to 
issues in model design or implementation. In order to 
improve the accuracy of theoretical models, it may be 
necessary to re-examine the assumptions, parameter 
settings, or data processing methods of the model to ensure 
that it better adapts to practical application scenarios. 
 
 
Figure 6 Accuracy prediction and testing results of 
models in three open datasets 

Strengthening Accounting Information Systems with Advanced Big… Informatica 49 (2025) 109–120 117 
In the experimental validation section of this article, 
we not only emphasized the significant improvement in 
model accuracy after feature selection, but also conducted 
more in-depth statistical analysis and validation to support 
our conclusions. Firstly, we compared the performance of 
the models before and after feature selection using the data 
in Table 2. From the table, it can be seen that the feature 
set after feature selection has improved accuracy by 
6.25%, while the training time has also been reduced. This 
result not only indicates that feature selection can reduce 
data complexity and improve model learning efficiency, 
but also validates our effectiveness and accuracy in the 
feature selection process. To further enhance the 
credibility of the conclusion, we conducted a statistical 
significance test. Specifically, we used paired sample t-
test to compare the difference in model accuracy before 
and after feature selection. The results indicate that this 
difference is statistically significant (p<0.05), further 
supporting our conclusion. In addition, we conducted a 
comprehensive comparison and analysis of the 
performance of different algorithms on three open 
datasets. As shown in Figure 6, although there is no 
significant difference between the results of the classical 
discrete model and the model proposed in this paper in 
some cases, we still emphasize the applicability of the 
classical model in specific environments or datasets. This 
viewpoint not only reflects our comprehensiveness and 
objectivity in model selection, but also provides readers 
with richer information and references.  
As shown in Figures 7 and 8. As the sample size 
increases, the training time also increases. This is because 
larger datasets require the model to perform more 
calculations and iterations to fit complex patterns and 
relationships in the data. Therefore, when dealing with 
large-scale datasets, it is necessary to consider the training 
efficiency of the model and the required computational 
resources. It is valuable to compare the accounting risk 
identification model in this article with the models in 
general traditional methods in terms of accuracy and 
average absolute error. This helps to evaluate the 
advantages and disadvantages of the model in practical 
applications, as well as its performance on different 
indicators. 
 
 
Figure 7: Accuracy comparison 
 
Through the case analysis of introducing the 
algorithm in this article into enterprise management 
accounting, it can be understood that intelligent 
accounting information systems can save costs and create 
more value for enterprises. Through continuous 
exploration in the field of management accounting 
informatization, multiple major businesses of enterprises 
have achieved first place in the industry, and the employee 
development index is far higher than the average of 
Baosteel Group. According to relevant professionals, an 
intelligent management accounting information platform 
is the future development trend of enterprises. In order to 
win in competition, enterprises must combine various IT 
technologies (big data, BI, artificial intelligence, cloud 
computing, etc.) to establish a management accounting 
information system that is in line with their own 
development. 
 
 
Figure 8: Comparison of F1 values for different 
algorithms 
 
According to the data in Figure 8, the recall rate and 
F1 value of our algorithm are higher than the other two 
algorithms. This result verifies that the algorithm has 
certain advantages compared to other algorithms. In the 
input factors of the model, factors such as actual situation 
are related to the accuracy of the prediction effect. 
Analyzing the prediction effect from multiple factors can 
improve prediction accuracy, and higher prediction 
accuracy can enable enterprises to effectively control 
financial risks. 
In this article, we propose and validate an 
optimization scheme for accounting information systems 
based on big data mining algorithms, which has achieved 
significant results in feature selection and model accuracy. 
By comparing the data accuracy and training time before 
and after feature selection, we found that the model after 
feature selection significantly reduced data complexity 
and training time while maintaining high accuracy. This 
discovery emphasizes the importance of feature selection 
in improving model performance and efficiency. 
Compared with classical discrete models and project 
response theory models, our model performs well in 
accuracy prediction and testing results on three open 
datasets. Although classical discrete models may still be 
applicable in some cases, our model demonstrates higher 
accuracy in most cases. This may be due to our model 

118 Informatica 49 (2025) 109–120 S. Yang 
adopting more advanced algorithms and a more 
comprehensive feature set, which can better capture 
complex patterns and relationships in the data. It should 
be noted that the accuracy of the theoretical model 
reflected in the project is the lowest, which may be due to 
certain issues in model design or implementation. In order 
to improve the accuracy of the model, it may be necessary 
to re-examine its assumptions, parameter settings, or data 
processing methods in the future to ensure that it better 
adapts to practical application scenarios. 
 
6 Conclusion 
This study delves into the optimization of accounting 
information systems based on artificial intelligence 
algorithms, especially in the context of the big data era. By 
introducing big data mining algorithms, accounting 
information systems have been significantly improved, 
effectively addressing the challenges of processing 
massive amounts of data. These algorithms not only 
improve data processing efficiency and reduce manual 
intervention, but also achieve automation in data cleaning, 
transformation, and analysis. These advances not only 
optimize the workflow of accounting practice, but also 
enhance the accuracy and efficiency of data processing. 
Experimental verification shows that compared with 
traditional single user authentication techniques. This 
means that in practical applications, accounting personnel 
can obtain and process data faster, improving work 
efficiency. Although this study conducted model 
recognition and data integrity verification in terms of data 
security and privacy protection, it did not delve into how 
to implement higher-level data protection and privacy 
encryption technologies at the algorithmic level. With the 
continuous development of big data and artificial 
intelligence technology, data security and privacy 
protection will become increasingly important issues. In 
response to the above limitations, future research can 
further explore the application of other artificial 
intelligence technologies (such as deep learning, 
reinforcement learning, etc.) in accounting information 
systems to find more suitable combinations of algorithms 
and technologies for specific scenarios, thereby further 
improving the performance and efficiency of the system. 
 
References 
[1] Tang, W., Yang, S., & Khishe, M. (2023). Profit 
prediction optimization using financial accounting 
information system by optimized DLSTM. Heliyon, 
9(9), 1. 
https://doi.org/10.1016/j.heliyon.2023.e19431 
[2] Bhima, B., Zahra, A. R. A., Nurtino, T., & Firli, M. Z. 
(2023). Enhancing organizational efficiency through 
the integration of artificial intelligence in 
management information systems. APTISI 
Transactions on Management, 7(3), 282-289. 
https://doi.org/10.33050/atm.v7i3.2146  
[3] Hu, K. H., Chen, F. H., Hsu, M. F., & Tzeng, G. H. 
(2021). Identifying key factors for adopting artificial 
intelligence-enabled auditing techniques by joint 
utilization of fuzzy-rough set theory and MRDM 
technique. Technological and Economic 
Development of Economy, 27(2), 459-492. 
https://doi.org/10.3846/TEDE.2020.13181 
[4] Berdiyeva, O., Islam, M. U., & Saeedi, M. (2021). 
Artificial intelligence in accounting and finance: 
Meta-analysis. International Business Review, 3(1), 
56-79. https://doi.org/10.37435/NBR21032502 
[5] Xia, W. H., Zhou, D., Xia, Q. Y., & Zhang, L. R. 
(2020). Design and implementation path of 
intelligent transportation information system based 
on artificial intelligence technology. The Journal of 
Engineering, 2020(13), 482-485. 
https://doi.org/10.1049/joe.2019.1196  
[7] Hua, H., Wei, Z., Qin, Y., Wang, T., Li, L., & Cao, J. 
(2021). Review of distributed control and 
optimization in energy internet: From traditional 
methods to artificial intelligence‐based methods. IET 
Cyber‐Physical Systems: Theory & Applications, 
6(2), 63-79. https://doi.org/10.1049/cps2.12007 
[9] Saragih, A. H., Reyhani, Q., Setyowati, M. S., & 
Hendrawan, A. (2023). The potential of an artificial 
intelligence (AI) application for the tax 
administration system’s modernization: the case of 
Indonesia. Artificial Intelligence and Law, 31(3), 
491-514. https://doi.org/10.1007/s10506-022-
09321-y 
[10] Faccia, A., & Petratos, P. (2021). Blockchain, 
enterprise resource planning (ERP) and accounting 
information systems (AIS): Research on e-
procurement and system integration. Applied 
Sciences, 11(15), 6792. 
https://doi.org/10.3390/app11156792 
[12] Andronie, M., Iatagan, M., Uță, C., Hurloiu, I., 
Dijmărescu, A., & Dijmărescu, I. (2023). Big data 
management algorithms in artificial Internet of 
Things-based fintech. Oeconomia Copernicana, 
14(3), 769-793. 
https://doi.org/10.1109/ISAIEE57420.2022.00032 
[13] Ahmad, A. (2024). Ethical implications of artificial 
intelligence in accounting: A framework for 
responsible ai adoption in multinational 
corporations in Jordan. International Journal of Data 
and Network Science, 8(1), 401-414. 
https://doi.org/10.5267/j.ijdns.2023.9.014 
[14] Peng, Y., Ahmad, S. F., Ahmad, A. Y. B., Al Shaikh, 
M. S., Daoud, M. K., & Alhamdi, F. M. H. (2023). 
Riding the waves of artificial intelligence in 
advancing accounting and its implications for 
sustainable development goals. Sustainability, 
15(19), 14165. https://doi.org/10.3390/su151914165 
[15] Ahmad, A. (2024). Ethical implications of artificial 
intelligence in accounting: A framework for 
responsible ai adoption in multinational 
corporations in Jordan. International Journal of Data 
and Network Science, 8(1), 401-414. 
https://doi.org/10.5267/j.ijdns.2023.9.014 
[16] Peng, Y., Ahmad, S. F., Ahmad, A. Y. B., Al Shaikh, 
M. S., Daoud, M. K., & Alhamdi, F. M. H. (2023). 
Riding the waves of artificial intelligence in 
advancing accounting and its implications for 

Strengthening Accounting Information Systems with Advanced Big… Informatica 49 (2025) 109–120 119 
sustainable development goals. Sustainability, 
15(19), 14165. https://doi.org/10.3390/su151914165 
[17] Zhang, Y., Xiong, F., Xie, Y., Fan, X., & Gu, H. 
(2020). The impact of artificial intelligence and 
blockchain on the accounting profession. IEEE 
Access, 8, 110461-110477. 
https://doi.org/10.1109/ACCESS.2020.3000505 
[18] Lee, C. S., & Tajudeen, F. P. (2020). Usage and 
impact of artificial intelligence on accounting: 
Evidence from Malaysian organisations. Asian 
Journal of Business and Accounting, 13(1), 
https://doi.org/10.22452/ajba.vol13no1.8 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

120 Informatica 49 (2025) 109–120 S. Yang