Analysis of Innovative Applications of Intelligent Technology in Economic Management and its Decision Support Mechanisms

The basic theory and method layer includes the basic theories and methods of strategy, economy, cost, performance management, and so on. Data mining is characterized by applicability and engineering, its applicability means that data mining is a combination of theoretical algorithms and application practice, and its engineering characteristic means that data mining is an engineering process composed of multiple steps. Data mining originates from the needs of practice and serves the practice at the same time. Therefore, the basic theories and methods such as strategic management and strategic management accounting are of guiding significance for carrying out data mining activities and are the fundamental part of the framework.

The data storage layer is the physical foundation of the overall architecture of the enterprise economic management system based on data mining technology, including various database systems of different business departments related to enterprise operation and management. These databases store all kinds of raw data related to economic management. Once the relevant information is properly prepared, it can provide the basis for the establishment of relevant models and data analysis.

Information analysis and integration is shown in Figure 2. In the process of informatization, the database of an enterprise contains various types of data, but its distribution is scattered and its structure is different, which is not suitable for direct data mining. The information analysis and integration layer contains a series of data preprocessing tools and methods to clean up, integrate, transform, and statute the original data. These preprocessing techniques are used before data mining and can effectively improve the quality of data mining.

Figure 2.

Information analysis and integration

The knowledge discovery layer is made up of a variety of data mining methods and tools and models for implementing them, which must be combined with the requirements of enterprise economic management. In the knowledge discovery layer, the transformation of “information to knowledge” is realized through the execution of data analysis and mining in related fields to provide support for enterprise economic management.

The cross-industry data mining process standard (DRISP-DM) emphasizes that data mining is not only the organization or presentation of data, nor is it just data analysis and statistical modeling, but is a complete process from understanding business needs, seeking solutions to being tested in practice. A complete data mining process is shown in Figure 3, and the whole process generally includes six stages: business understanding, data understanding, data preparation, modeling, evaluation, and deployment.

Figure 3.

General process of data mining

On the basis of the principles and general architecture of data mining applied to enterprise economic management in the previous section, the implementation process of data mining applied to enterprise economic management is proposed:

1) Firstly, based on the needs of enterprise economic management, define the problems to be analyzed.

In order to realize the risk management of economic data by the enterprise economic management system, it is necessary to combine the Apriori association rule algorithm in big data to mine economic risk data [25]. The specific realization steps are as follows:

Assume that the total set of enterprise economic data is I = {i₁,i₂,…,i_n} and the set of risk data is D, where T represents a subset composed of each risk data. Assuming that X and Y are two events in the subset of risk data, the Apriori association rule can be constituted, and the specific expression is: (1) T:X⇒Y

In order to assess the value of association rules, two indicators, confidence and support, are generally used for measurement [26]. Among them, the confidence degree indicates the size of the probability of event Y occurring under the premise of event X occurring, and the specific expression is: (2) Confidence(X⇒Y)=P(X∪Y)P(X) where P(X ∪ Y) is the number of data subsets containing both event X and event Y, and P(X) is the probability of event X occurring alone.

The support represents the probability of event X occurring simultaneously with event Y , which is expressed as: (3) Support(X,Y)=P(X∪Y)| D |

Where |D| is the total number of elements in the economic risk data set. In the enterprise economic management system, the confidence level and support level of economic risk data are high. Therefore, in order to mine these risk data, it is necessary to set the minimum thresholds of confidence and support, which are defined as con_min and sup_min, respectively. When the support and confidence of a certain data is greater than the minimum threshold indicator, it means that the data conforms to the strong correlation rule, which is that it is economic risk data.Similarly, if the support and confidence of the data are within the minimum threshold, it means that the data does not belong to the economic risk data and does not need to be extracted.

Through the above steps, the extraction of risk data from an enterprise’s economic management system can be realized through the use of confidence and support levels to provide data support for subsequent economic strategy decisions.

Budget management is the core link in the economic management of enterprises, and the budget management subsystem of enterprises is shown in Figure 4. It is used to assist with enterprise planning for various work operations, and provides an evaluation basis for relevant operations based on budgetary results, which in turn points to the strategic direction of the enterprise’s digital reform.

Figure 4.

Budget management subsystem

Faced with the volatile market environment, enterprises can make timely adjustments with the help of intelligent budget management. It is based on data accounting and basic decision-making work, in order to help enterprises digitalize further transformation, adapt to the strategic objectives of enterprises, and provide support for management to make the right decisions.

The architecture of the enterprise’s performance management subsystem is shown in Figure 5. The relevant management can use the performance management information system to understand the internal performance information of the enterprise. Based on this, the overall strategic direction of the enterprise and performance objectives are determined. Each department through the information management system to transmit instructions to the same time, according to the system according to the characteristics of each employee’s work, the appropriate and effective performance indicators, the company’s strategy and each person linked.

Figure 5.

Performance management subsystem architecture

By applying big data technology to performance management methods such as balanced scorecard, managers can quickly and effectively access and analyze key performance evaluation indicators such as finance, sales, customers, etc., and understand the performance management of the enterprise in real time through big data technology. With the help of big data technology, data transfer can be realized and made compatible with different enterprise systems to solve the problem of efficient information transfer.Performance management can be monitored beforehand and by employees, so that the company’s business activities are truly aligned with its vision.

The cost management system is an important part of the digital transformation of the enterprise. The cost management subsystem architecture is shown in Figure 6. Based on the characteristics of the enterprise, the functional requirements of the cost management subsystem lie in cost aggregation, expense allocation, and running the cost-method. On this basis, through the cost management subsystem, the enterprise uses the system to analyze the composition of the enterprise’s investment costs and the proportion of each part with the help of big data technology, so as to more accurately identify the reasonable cost drivers of the enterprise, and build a suitable job cost pool based on this.

Figure 6.

Cost management subsystem architecture

Enterprise economic management in order to assess the economic management risk, need to establish a mathematical model to achieve the prediction of the risk value of economic data. Assuming that the original sequence of enterprise economic risk data is X₀, the specific expression is: (4) X0=(x10,x20,⋯,xn0)

The original sequence information is combined and the final new sequence expression is generated: (5) X1=(x11,∑n=12x21,⋯,∑n=1nxn1)

The combination of Eq. 4 and Eq. 5 can be used to obtain the basic model of enterprise economic risk prediction, and the specific model expression is: (6) βi=αxi1+xi0

where α and β are model parameters, respectively. By discretizing the original sequence of enterprise economic risk, the time response sequence about time t can be obtained, and the specific expression is: (7) X(k+1)=(X0−βα)βα

where k is a natural number with a value range between [1,n]. By accumulating the above time response series, thus obtaining the enterprise economic risk assessment model, the specific expression is: (8) X(k+1)=ωX1(k+1),0≤ω≤1

Where ω is the predicted value of enterprise economic risk weight. By calculating the weight prediction value and combining it with the risk indicators obtained by mining association rule algorithm, the risk assessment of enterprise economic management can be realized.

Combining the system part with the algorithm, the design of enterprise economic management risk assessment subsystem based on association rule algorithm is completed.

The operation mechanism of the enterprise’s decision support subsystem is shown in Figure 7, based on the economic management system’s analysis of the enterprise’s internal and external historical data, thus obtaining reports with decision-making assistance to assist the enterprise’s decision makers in making decisions, and ultimately reducing enterprise costs and improving investment efficiency. Through the economic management system, the enterprise connects all modules within the enterprise, which strengthens the flow of information and makes the company’s decision-making more accurate. The decision support subsystem summarizes both internal and external information of the enterprise, while also connecting information exchange among internal systems to obtain diversified information for the company’s decision-making.

Figure 7.

Enterprise decision support subsystem operation mechanism

The major subsystems of the intelligent economic management system of the enterprise are closely linked to each other and work together to make it efficient, thus accelerating the pace of digital transformation.Each subsystem obtains information through the information management system and performs information exchange. The cost management system saves the enterprise investment cost to a certain extent by implementing the job cost method, and feeds the data back to the budget management system as much as possible to provide the basis for the next stage of budget. At the same time, the budget management system also feeds data back to the cost management system’s budget, which provides the scope definition for the cost management system’s operation. The budget management system and the performance management system assist each other, the budget can assist the enterprise to improve performance, the performance level is also one of the reference indicators of the budget management system, the performance management system is also involved in the performance evaluation and appraisal of each system. The decision support system aggregates information from various sub-systems to form the enterprise’s unique decision-making data, which forms the basis for the digital transformation of the enterprise’s entire intelligent economic management process.

Founded in 1995, P Company is a leading enterprise in China’s investment, from a local small business on the verge of bankruptcy to the growth of the current internationalization of large enterprises, P Company has seized every opportunity for development, in line with the trend of the times, and now from the traditional investment enterprises, the transformation of the whole society to incubate high-tech industries on the angel financing platform. It is committed to become an Internet enterprise, it is to change the traditional investment system, from a closed system into a link in the network interconnection, through this node to open up all aspects of the resources, to create a new platform for win-win situation, to achieve the win-win value-added resources.

Since the establishment of P Company, it has insisted on putting users’ needs in the first place, and has constantly pursued management innovation, and innovation has driven the healthy growth and development of the enterprise.The changes in operating income and net profit from 2009 to 2023 are shown in Figure 8, and P Company realized operating income of 179.675 billion yuan in the whole year of 2023, with a year-on-year growth of 26.54%.P Company 2023 realized a profit of 6.052 billion yuan, and net profit reached a record high. Net profit of 6.052 billion yuan, net profit hit a record high, P companies 2023 all categories of market share are showing an increase in the trend, and continued to expand in 2018.

Figure 8.

Change of operating income and net profit from 2009 to 2023

P Company not only has excellent performance, but also innovative economic management, which has made its corporate culture well-known both at home and abroad.In 2012, P Company was awarded the CIMA Management Practice of the Year Award at the CIMA Annual General Meeting for its ‘human-monitoring’ economic management accounting model.

With the continuous expansion of the scale of P enterprise, the methods of its internal management are becoming increasingly inapplicable, which greatly reduces the management efficiency of the enterprise.At the same time, as times change, the concept of management should also keep pace with them.The traditional enterprise management model has a hierarchical concept that is not conducive to young employees realizing their value.

In the stage of diversification strategy, with the intelligent technical support of big data, cloud computing, business intelligence, etc., Company P applies data mining technology in economic management.Company P builds an economic management system based on data mining technology around the application of data mining technology in enterprise economic strategy management.

The construction of the system framework aims to achieve the following goals for Company P:

1) Explore effective ways for enterprises to make full use of information in the era of knowledge economy and deepen their understanding of data and information.

The system sometimes appears to perform repetitive computational behavior in the economic data mining operation, which is caused by the decrease of the support threshold.Therefore, this mining algorithm used in this paper is built on the basis of already available mining information, and its resulting frequent items are also subordinated to a new support threshold.Its saving method is a Hash structure, which also enables the retrieval of the support count of the frequent itemset, resulting in significantly higher overall mining efficiency.

In the first calculation operation of the frequent itemset, the Apriori association rule algorithm is applied to save the resulting hierarchical data of the frequent itemset into the targeted thresholds.

Assume that LOrg₁, LNew₁ exist in all pattern types of Lk₁, Lk₂ respectively, and let Lk₃ be composed of a non-empty subset of LOrg₁, LNew₁ spliced together.

The IUA_GenL_k function obtains C_k¹, C_k² and C_k³ through the self-connection of L_k-1¹ and L_k-1², respectively, and the algorithm is stopped as soon as the null value of L_k occurs, through the continuous screening of the combinations C_kⁱ and L_k , which will eventually appear at the threshold of the new support. At this point, the association rules obtained for the economic indicators have a mutual correspondence with their respective support levels. For a single economic indicator, the hierarchical data involved are usually called descriptive items.

This paper focuses on three association rule algorithms, IUA, HIUA, and Apriori, and explores their performance in terms of runtime.C# (Net 4.0) provides support for the implementation of the above algorithms on an i7 processor, 16G RAM, and Win10 system. The update operation for frequent pattern sets does not apply to two cases, one is when there is an increasing trend in the support threshold and the other is when there is a fluctuating change in the confidence threshold. Based on this situation, it can be intuitively found that at this time the speed advantage that IUA and HIUA have is more obvious.

In this paper, we only consider the situation when there is a downward trend in the support degree, and compare and analyze the size of the time used by each algorithm in the calculation of frequent pattern sets, and the specific running time is shown in Figure 9. The sample data involved are derived from the economic data of P enterprise from 2009 to 2023, which can reflect the economic indicators of P enterprise in the period of 2009-2023, involving 112052 records, while the number of financial indicators involved in each record is 16. Among them, the X and Y coordinates represent different meanings, for the X coordinate, its support threshold ranges from 0.2 to 0.3, and is measured in 0.01 steps. For the Y-coordinate, it mainly reflects the difference in the running time of the three algorithms. It can be seen that the Apriori association rule used in this paper achieves optimal pairwise performance at different support thresholds.

Figure 9.

Comparison of operation performance of the three algorithms

Selecting appropriate economic indicators is the key to analyze the success of enterprise economic risk management, this study based on business needs, using association rules for mining a total of 16 economic indicators were collected, respectively, gross profit margin a, net profit margin b, return on net assets c, basic earnings per share d, return on total assets e, quick ratio f, current ratio g, cash ratio h, gearing ratio i, interest coverage multiple j, the Total asset turnover k, accounts receivable turnover l, inventory turnover m, total asset growth rate n, net profit growth rate o, operating income growth rate p. There is a certain degree of correlation between these financial indicators, run the SAS software through the correlation coefficient calculation module, to get the correlation coefficient between the indicators as shown in Figure 10. The indicators with positive correlation (P) and very high negative correlation between the indicators are eliminated, as well as several groups of indicators with higher absolute values of correlation coefficients in the economic indicators under the same module. Finally, 11 economic risk indicators are mined, which are gross profit margin a, net profit margin b, return on net assets c, basic earnings per share d, current ratio g, cash ratio h, gearing ratio i, accounts receivable turnover l, inventory turnover m, total assets growth rate n, and net profit growth rate o.

Figure 10.

Correlation coefficient between economic indicators

It is worth noting that the number of rules varies with the confidence and support thresholds.The variation in the number of correlation rules for the analysis of economic risk indicators from 2009 to 2023 for Company P is shown in Figure 11, where the X, Y, and Z coordinates denote the support, confidence, and number of rules, respectively.

Figure 11.

Changes in the number of association rules for economic risk indicators

Enterprise economic data has continuity, while data mined using association rules is generally discrete. Therefore, it is necessary to convert continuous data into discrete data by defining economic data indicators, so that each economic data can be divided into risk levels as shown in Table 1. 11 economic indicators, a total of 5 risk levels, from level 1 to level 5, level 5 indicates the highest level of risk, which urgently requires enterprise management to make urgent decisions.

Table 2 shows the number of association rules for the economic risks of Company P from 2009 to 2023, where the Apriori association rule algorithm uses different support and confidence thresholds. It can be intuitively reflected that when the support threshold is taken as 0.6 and the confidence threshold is taken as 0.3, there are as many as 84 economic management risks involved in the operation process of Company P from 2009 to 2023.