Financial performance of intelligent manufacturing enterprises based on fuzzy neural network and data twinning

Abstract

As we enter the 21st century, with the rapid development of information technology and advanced manufacturing technology and the rapid development of intelligent manufacturing in China, the national investment in intelligent manufacturing research has increased greatly. And in this field to establish research projects; our intelligent manufacturing has been in-depth development of the moment. Assessing the financial performance of a company is to assess the profitability, sustainability and growth ability of a listed company and plays an important role in reducing investment risk, safety and security. This document assesses the financial performance of intelligent manufacturing enterprises by studying and analyzing the improvement of financial management model in intelligent manufacturing model and the optimization of cost involved. Based on fuzzy neural network and data generation technology, this helps to improve the competitive advantage of enterprises and achieve sustainable development and change.

Keywords

Fuzzy neural network data twinning intelligent manufacturing enterprise financial performance evaluation

1 Introduction

In recent years, China’s manufacturing industry has developed from traditional to modern, and major changes have taken place in this sector. The combination of intelligent manufacturing, computerization and industrialization innovation has attracted great attention from all countries and people. The development of intelligent manufacturing meets the needs of the development of manufacturing sector in China. It is an important step to improve our country’s economic ability. Therefore, when we entered the 21st century, with the rapid development of information technology and advanced manufacturing technology, and the rapid development of intelligent manufacturing industry in China, China’s investment in intelligent manufacturing research has increased greatly. And in this field to establish research projects, intelligent manufacturing in China has been in-depth development. Today, the developed countries are speeding up the process of revitalizing the manufacturing industry, while the developing countries are taking over industrial transfer at a lower cost, and our country is facing “double pressure”. Therefore, the pace of research and development of intelligent manufacturing technology needs to be accelerated to deal with the erosion of low-cost advantages. The adoption of advanced intelligent manufacturing technology and new and more environmentally friendly energy saving equipment can also fundamentallysolve the problem of reducing energy efficiency in China’s traditional manufacturing industry.

Intelligent product manufacturing, intelligent production and intelligent services supported by tangible network systems have become an important choice for countries in restructuring global value chains and restructuring the International Finance and Development Division. The production mode of intelligent manufacturing will have a deeper impact on the organization and management of enterprise cost, as well as manufacturing technology, daily operation and so on. This in turn leads to changes in the financial structure. assessing the financial performance of a company is assessing the profitability, sustainability and growth ability of a listed company and plays an important role in reducing investment risk. financial security and profit expectations this document assesses the financial performance of an intelligent manufacturing enterprise by studying and analyzing improvements to the financial management model in the smart manufacturing model and optimizing the costs involved. Based on fuzzy neural network and dual data generation technology, this helps to improve the competitive advantage of enterprises and achieve sustainable development and change.

2 Related work

According to the literature [1] the essence of intelligent manufacturing is to use the basic concepts and methods of artificial intelligence to solve the problems in the actual manufacturing process. Literature [2] the concept of intelligent manufacturing system has been developed, pointing out that this is a flexible manufacturing system, which is combined with manual. According to [3] literature, intelligent manufacturing is a very intelligent and computerized combination of manufacturing industry, which can predict, analyze and solve life production problems. Literature [4] analysis of the promulgation of “China Intelligence system 2025” summarizes the challenges faced by China in the field of intelligent manufacturing, that is, it is necessary to improve the environment of manufacturing development and improve production efficiency. But it basically does not answer the “where” and “how” questions of Chinese manufacturing [5]. Literature points out that intelligent manufacturing industry is an important technology for China’s industrial development. In order to solve our backward problem in developing intelligent manufacturing, the government has formulated a two-stage development strategy. The first envisaged expanding the scope of digitalization of manufacturing within and among enterprises and its practical application among enterprises; the second envisaged the introduction of widespread application of computerization of manufacturing. According to the literature [6], knowledge manufacturing is in the early stage of development, and there are still differences with the situation of developed countries.

As for evaluating the financial performance of the company, foreign experts began to study the financial performance of the company as early as the beginning of the last century, and DuPont’s Dolsonbrown company developed a research-based financial analysis system for the company. Called Dupont financial analysis system, the system is mainly based on the return rate of net assets of the company system and is used as a financial performance survey method. The basis of decomposition is the factors that affect the ability of the enterprise, and analyze the change of the net assets of the enterprise and the change of different decomposition indexes, so as to analyze the factors that affect the ability of the enterprise and study the reasons in depth. Document [7] Based on the theory and knowledge of applicable financial, accounting, economic and management disciplines, a comprehensive analysis of the financial results system of internal management of the company is carried out, and the results of the research on the financial performance of western countries over the past 100 years are streamlined and summarized. This is a performance appraisal and academic research. Analysis of enterprise performance evaluation [8] literature shows that concept and structure are very important to guide enterprise performance evaluation, and enterprises can change and adjust their status and development strategy according to standardized model. At the literature [9] meeting, the importance of financial performance evaluation in business management is discussed, and it is pointed out that enterprises can find out the defects in the process of performance evaluation in time, correct the defects in time, and promote the sustainable development of enterprises. Document [10] report summarizes the difficulties and shortcomings of enterprises in adopting traditional financial performance evaluation methods, and puts forward some suggestions to improve the financial performance evaluation system through specific case analysis. Literature [11] This study focuses on 40 enterprises, calculates financial data related to enterprises, analyzes the factors that affect the development of enterprises, and infer the relationship between enterprise development and capital structure, business scale and corporate social responsibility.

In general, national researchers do not pay enough attention to the theory of financial performance assessment itself, which is often based on foreign research theories and combines the establishment of evaluation systems with national reality and empirical analysis. evaluation systems are designed for research purposes and have no practical value for smart manufacturing companies; although smart manufacturing companies have indications that they have no value. The importance of the objectives of the Ministry of Finance is gradually recognized, but the relevant criteria for selecting non-financial indicators have not been determined, and if enterprises choose more indicators, the cost of evaluation will increase, which requires enterprises to invest a lot of material and human resources. In addition, no quantitative criteria for non-financial indicators have been established, so a more comprehensive evaluation system has been established. Results-based assessments are often not objective; many country studies are conducted from a macroeconomic perspective, analysing the performance of the processing industry, and relatively few in-depth analysis and research on the financial performance evaluation of specific types of companies or enterprises. Therefore, recommendations based on these recommendations play a limited guiding role in assessing the performance of smart manufacturing enterprises.

3 Analysis on financial performance evaluation of intelligent manufacturing enterprises

3.1 Evaluation model construction

Using existing resources and data in the business process can help enterprises improve their conceptual framework and improve the structure and knowledge accuracy of enterprise analysis. it should be noted that the AHP method is not perfect, and there are some shortcomings in its application. therefore, this paper ensures the objectivity of the data by combining quantitative and qualitative, while improving the practicability and legitimacy of the research results. In calculating the final results, a broad and vague assessment method was used to make it more accurate, and a comprehensive review of the financial performance assessment method used not only quality indicators, but also quality indicators. The difficulties encountered in developing possible evaluation criteria should be compared with quantitative indicators. Although more accurate values can be obtained, more scholars have proposed that enterprise comparison methods with historical benchmarks and comparisons can be used. But in practice, the development of different enterprises is different; quality indicators can not get specific value. This will reduce the accuracy of the evaluation results. The above analysis shows that the characteristics of the financial performance evaluation process are relatively vague, highly complex and multifaceted. basic consequences, this document uses the fuzzy neural network evaluation model and the AHP evaluation model. the aim is to show more objectively and concretely the quantitative and qualitative indicators studied and to improve the authenticity, reliability and objectivity of the evaluation results.

This document is mainly divided into three stages to develop the evaluation model.

The first step is to establish an evaluation index system and classification model to classify the indicators.

The second step is to establish a diagnostic matrix and classify the importance of the index by measuring the elements of the matrix.

A typical value of different elements is calculated to obtain the weight of different indexes, and the consistency of the matrix is tested after calculating the typical maximum value of the Lambda Max. $λ max j = \frac{1}{m} \sum_{i = 1}^{m} \frac{(Aw)_{i}}{w_{i}}$ (1) $C \cdot I = \frac{λ max - m}{m - 1}$ (2) $C \cdot R = \frac{C \cdot I}{R \cdot I}$ (3)

C.I. is the t CR < 0.10 standard, the calculation result can adopt consistency standard, but not consistency, which needs to be recalculated or adjusted.

Finally, the data is processed by comprehensive evaluation of neural networks. First, establish a set of commentaries to translate different calculations into different results. Second, enterprises should specify their own securities, quality indicators should be determined by expert scoring method, and quantitative indicators should be calculated by fuzzy function calculation method. Third, the weights of different indexes of ranking submatrix are processed by a fuzzy transformation method and finally evaluated.

3.2 Construction of evaluation index system

In this study, the author based on the relevant knowledge theory, and to the greatest extent based on theory and practice, and established a system of evaluation indicators of results. This paper collects more data by distributing questionnaires, and this study reviews a large amount of literature to finalize the results of previous studies. Obtain appropriate financial assessment indicators for theoretical and methodological analysis of these indicators. This document selects 50 specific indicators based on relevant theoretical and developmental characteristics and raises questions in the questionnaire. These indicators have been incorporated and a significant amount of data has been collected. By combining the collected data and eliminating the substandard indicators in time, a relatively sound financial performance evaluation index system has been established, see Tables 1 and 2 for details.

Table 1
Category I indicators

Indicators Mean Standard deviation

Net sales rate 6.364 1.129

Net profit growth rate 6.961 1.134

Gross operating margin 6.286 1.410

Current ratio 5.674 1.294

Asset-liability ratio 5.948 1.348

Inventory turnover 5.935 1.253

Sales cash ratio 5.247 1.310

Cost-cost margins 5.718 1.345

Turnover of accounts receivable 5.438 1.481

Cash flow liabilities ratio 5.483 1.385

Quick ratio 5.531 1.432

Growth rate of main business 5.462 1.398

Gross operating margin 6.286 1.241

Current ratio 5.874 1.284

Market share 5.948 1.343

Capital preservation and appreciation rate 5.481 1.531

Indicators	Mean	Standard deviation
Net sales rate	6.364	1.129
Net profit growth rate	6.961	1.134
Gross operating margin	6.286	1.410
Current ratio	5.674	1.294
Asset-liability ratio	5.948	1.348
Inventory turnover	5.935	1.253
Sales cash ratio	5.247	1.310
Cost-cost margins	5.718	1.345
Turnover of accounts receivable	5.438	1.481
Cash flow liabilities ratio	5.483	1.385
Quick ratio	5.531	1.432
Growth rate of main business	5.462	1.398
Gross operating margin	6.286	1.241
Current ratio	5.874	1.284
Market share	5.948	1.343
Capital preservation and appreciation rate	5.481	1.531

Table 2

Category II indicators

Indicators	Mean	Standard deviation
Multiple interest guarantee	4.356	1.8362
Business cycle	2.484	1.9288
Total assets turnover	3.470	1.9335
Cash operating index	2.040	1.9134
Number of patent applications	3.126	2.295
Customer satisfaction	3.013	1.8231
Brand awareness	2.530	1.9303
Knowledge Workers Ratio	2.042	1.9134
Collaborative Innovation Management Level	4.357	2.0832
Information system update	2.439	1.9467
R &D investment intensity	1.984	1.7929
Staff training	2.042	1.8159
Labour productivity	2.544	1.8493
Staff motivation	2.672	1.9378

As can be seen from Figs. 1 and 2 above, the analysis of the data in Table 2 shows that the difference and standard average of this index is high and the polarization is intensified. Therefore, on this basis, a new survey was carried out and new indicators were introduced. As shown in survey 2, it can be seen that the standard of indicators in the table is not different and the values are very similar, which indicates that the views on these indicators are relatively consistent. The above analysis of respondents shows that this indicator is one of the issues that enterprises need to address, and the importance of this issue is recognized. indicator comparison shows that there are some conceptual similarities between the total utilization rate of assets and the activity rate, liquidity rate and speed. therefore, the data in the tables of these two aspects should not be repeated to show that the training level and motivation of employees are low, which means that there is a greater internal gap in the human resource level of S companies. Develop professional and basic skills of employees and incorporate relevant human resources data into the corporate performance appraisal system for analytical evaluation.

Fig. 1

More significant(%).

Fig. 2

Importance and proportion.

According to the in-depth study and research of the literature,25 tertiary indicators (5 quality indicators and 20 quantitative indicators) were selected as the evaluation criteria, taking into account the current situation of enterprise development. As shown in Table 3, quality indicators are more easily available and can be studied through questionnaires, while quantitative indicators need to be calculated to obtain information.

Table 3

S Corporate Financial Performance Evaluation Index System

Target layer	Level I indicators	Secondary indicators	Level 3 indicators
S A of Financial Performance Evaluation	A1 of realistic financial performance	A11 profitability	A111 on return on net assets
			Gross margin A112 of main business
			A113 of cash ratio for sales
			A114 of cost-cost margins
			A115 of Cash Operating Index
		A12 solvency	A121 current ratio
			A122 of cash flow liabilities
			A123 of multiple interest guarantee
			A124 of asset-liability ratio
		Capacity A13 for development	A131 of net profit growth
			A132 growth rate of main business
			A133 of capital preservation and appreciation
		A14 of operational capacity	A141 of inventory turnover
			A142 of accounts receivable turnover
			A143 in total assets turnover
	A2 of potential development capacity	A21 of external market capacity	A211 market share
			A212 of brand awareness
			A213 of customer satisfaction
		A22 of internal management capacity	A221 of updates to information systems
			A222 human resources management level
			A223 labour
		A23 of innovative capacity	A231 knowledge worker ratio
			A232 of R &D investment intensity
			Number A233 patent applications
			A234 of Collaborative Innovation Management Level

By weight calculation, we can know the relevant data indicators of S company, the specific indicators are shown in the following Table 4:

Table 4

S Weight of corporate financial performance evaluation indicators

Target layer	Level I indicators	Weight (%)	Secondary indicators	Weight (%)	Level 3 indicators	Weight (%)
S A of Financial Performance Evaluation	A1 of realistic financial performance	75.00	A11 profitability	57.23	A111 on return on net assets	40.57
					Gross margin A112 of main business	25.60
					A113 of cash ratio for sales	16.50
					A114 of cost-cost margins	10.63
					A115 of Cash Operating Index	6.71
			A12 solvency	20.90	A121 current ratio	8.15
					A122 of cash flow liabilities	23.73
					A123 of multiple interest guarantee	15.61
					A124 of asset-liability ratio	52.52
			Capacity A13 for development	10.94	A131 of net profit growth	24.93
					A132 growth rate of main business	59.36
					A133 of capital preservation and appreciation	15.71
			A14 of operational capacity	10.94	A141 of inventory turnover	16.27
					A142 of accounts receivable turnover	49.37
					A143 in total assets turnover	34.36
	A2 of potential development capacity	25.00	A21 of external market capacity	54.99	A211 market share	26.93
					A212 of brand awareness	18.36
					A213 of customer satisfaction	5.90
			A22 of internal management capacity	24.02	A221 of updates to information systems	48.82
					A222 human resources management level	55.18
					A223 labour	25.05
			A23 of innovative capacity	20.98	A231 knowledge worker ratio	19.78
					A232 of R &D investment intensity	24.02
					Number A233 patent applications	20.98
					A234 of Collaborative Innovation Management Level	54.99

3.3 Analysis of comprehensive evaluation of fuzzy neural networks

First, the results of the attribution criteria and quantitative indicators are derived from the calculation formula, as shown in Table 5, in which the interest guarantee multiplier is usually not calculated, Because the company report does not provide information on the following: according to the definition and content of the multiple of security interests, the value in the research and analysis process is unlimited in this paper. An internal statistical report was extracted from the area of disclosure of S.A. financial statements after careful collection, review and analysis of relevant documents, see Table 5 for details.

Table 5
Numerical results of quantitative indicators and criteria for judging membership

Secondary indicators Level 3 indicators Indicator attributes Lower value Indicator values Benchmark value

A11 profitability A111 on return on net assets Positive –5.30% 12.35% 17%

Gross margin A112 of main business Positive –3.40% 24.09% 26.10%

A113 of cash ratio for sales Positive –4.00% 7.63% 9.70%

A114 of cost-cost margins Positive 0.00% 12.45% 23

A115 of Cash Operating Index Positive 0.43 0.78 1

A12 solvency A121 current ratio Positive –0.38% 44.29% 1

A122 of cash flow liabilities Positive 1 2.46 4.35

A123 of multiple interest guarantee Positive 3.6 ∞ 9.4

A124 of asset-liability ratio Reverse reverse 87% 33.92% 33.05%

A13 of operational capacity A141 of inventory turnover Positive 94.40% 112.16% 116%

A142 of accounts receivable turnover Positive –8.6% 7.36% 12.15%

A143 in total assets turnover Positive 0.00% 0.76% 33.57%

Capacity A14 for development A131 of net profit growth Positive 1.4 3.67 5

A132 growth rate of main business Positive 1.5 5.25 10.6

A133 of capital preservation and appreciation Positive 0.3 1.20 1.4

A21 of external market capacity A211 relative market share Positive 0.13% 14.23% 33.3%

A22 of internal management capacity A222 labour Positive 45468.2 765656.23 1243653.31

A23 of innovative capacity A232 of R &D investment intensity Positive 4.50% 18.81% 21.76%

Number A233 patent applications Positive 1.30% 4.43% 6.45%

A234 of Collaborative Innovation Management Level Positive 0 82 282

Secondary indicators	Level 3 indicators	Indicator attributes	Lower value	Indicator values	Benchmark value
A11 profitability	A111 on return on net assets	Positive	–5.30%	12.35%	17%
	Gross margin A112 of main business	Positive	–3.40%	24.09%	26.10%
	A113 of cash ratio for sales	Positive	–4.00%	7.63%	9.70%
	A114 of cost-cost margins	Positive	0.00%	12.45%	23
	A115 of Cash Operating Index	Positive	0.43	0.78	1
A12 solvency	A121 current ratio	Positive	–0.38%	44.29%	1
	A122 of cash flow liabilities	Positive	1	2.46	4.35
	A123 of multiple interest guarantee	Positive	3.6	∞	9.4
	A124 of asset-liability ratio	Reverse reverse	87%	33.92%	33.05%
A13 of operational capacity	A141 of inventory turnover	Positive	94.40%	112.16%	116%
	A142 of accounts receivable turnover	Positive	–8.6%	7.36%	12.15%
	A143 in total assets turnover	Positive	0.00%	0.76%	33.57%
Capacity A14 for development	A131 of net profit growth	Positive	1.4	3.67	5
	A132 growth rate of main business	Positive	1.5	5.25	10.6
	A133 of capital preservation and appreciation	Positive	0.3	1.20	1.4
A21 of external market capacity	A211 relative market share	Positive	0.13%	14.23%	33.3%
A22 of internal management capacity	A222 labour	Positive	45468.2	765656.23	1243653.31
A23 of innovative capacity	A232 of R &D investment intensity	Positive	4.50%	18.81%	21.76%
	Number A233 patent applications	Positive	1.30%	4.43%	6.45%
	A234 of Collaborative Innovation Management Level	Positive	0	82	282

Step 2, Based on the data in the table above, Take a certain calculation, Using relevant formulas, Determine the process of calculating the membership value of the enterprise. For positive indicators, Ensure that its minimum value is controlled at a, Equalize the distance within the b], Add three d2,d3,d4, points d1 = b, Order d2 = 3(b-a)/4 + a, d3 = (b-a)/2 + a, d4 = (b-a)/4 + a, d5 = b. Find out the comment set U = 1, 2,3, 4,5 = excellent, Good, In general, Poor, Difference corresponding index standard value: d1, d2, d3, d4, d5. For the relevant negative indicators, and shall be monitored within [B,A], which should include three values. the same calculation results as above are listed in the Table 6 below.

Table 6

Numerical distribution of quantitative indicators

Level 3 indicators	D5	X	D4	D3	X	D2	X	D1
A111 on return on net assets	–0.041	0.006	0.010	0.056	0.127	0.149	0.261	0.230
Gross margin A112 of main business	–0.033		0.034	0.132		0.212		0.291
A113 of cash ratio for sales	0.000		0.049	0.110	0.117	0.140	0.074	0.210
A114 of cost-cost margins	–0.030		–0.006	0.040		0.067		0.089
A115 of Cash Operating Index	0.420		0.576	0.752	0.790	0.836		1.100
A121 current ratio	1.000		1.756	2.350		3.925		4.070
A122 of cash flow liabilities	–0.300		–0.049	0.290	0.469	0.690		1.100
A123 of multiple interest guarantee	3.600		5.330	6.500		8.340	0.342	9.600
A124 of asset-liability ratio	0.840		0.742	0.562		0.44		0.331
A131 of net profit growth	0.000	0.006	0.086	0.196		0.245		0.349
A132 growth rate of main business	–0.087		–0.044	0.017		0.080	0.076	0.131
A133 of capital preservation and appreciation	0.945		1.018	1.042		1.161	1.121	1.150
A141 of inventory turnover	1.690		2.255	3.530	3.870	4.157		5.100
A142 of accounts receivable turnover	1.780		4.257	6.570		9.252		11.600
A143 in total assets turnover	0.180		0.540	0.910		1.270	1.300	1.500
A211 market share	0.001		0.063	0.176		0.258		0.340
A221 labour	48635.200		341714.728	631694.255	763446.230	930783.783		1235563.310
A231 knowledge worker ratio	0.043		0.081	0.183	0.182	0.273		0.300
A232 of R &D investment intensity	0.012		0.036	0.029	0.043	0.041		0.074

The third step is to calculate the membership value of the enterprise based on the above data and according to the relevant formula of trigonometric function. The following table shows the results of the calculation, see Table 7 for details.

Table 7

Membership of quantitative indicators

Level 3 indicators	U1	U2	U3	U4	U5
A111 on return on net assets	0	0.845	0.03	0	0
Gross margin A112 of main business	0.75	0.24	0	0	0
A113 of cash ratio for sales	0	0.42	0.69	0	0
A114 of cost-cost margins	0.26	0.74	0	0
A115 of Cash Operating Index	0	0.67	0.47	0	0
A121 current ratio	0	0	0.92	0.07	0
A122 of cash flow liabilities	0	0.58	0.53	0	0
A123 of multiple interest guarantee	0	0	0	0	0
A124 of asset-liability ratio	0.89	0.03	0	0	0
A131 of net profit growth	0	0	0	0.08	0.93
A132 growth rate of main business	0.13	0.86	0	0	0
A133 of capital preservation and appreciation	0.08	0.92	0	0	0
A141 of inventory turnover	0	0.49	0.35	0	0
A142 of accounts receivable turnover	0	0	0.63	0.42	0
A143 in total assets turnover	0.13	0.68	0	0	0
A211 market share	0	0	0.85	0.19	0
A221 labour	0	0.34	0.59	0	0
A231 knowledge worker ratio	0	0.15	0.89	0	0
A232 of R &D investment intensity	0	0	0.34	0.68	0
Number A233 patent applications	0	0	0	0.25	0.79

4 Design of fuzzy neural network based on data twinning

By solving a small set of problems, two nonparallel superclassifications are studied, but SVM contain reverse matrix operations, which require a large amount of memory and computational resources to process a large number of data sets. For the framework of neural network, this paper uses the concept of non-parallel boundary to realize the structure of neural network based on a two-pronged strategy.

Figure 1 shows the structure diagram of the tertiary neural network (foam-layered-output layer), which contains X input vectors, neural resolution (X) for X conversion to ramp, processed seal layer, and output layer learned from a subclass. samples of the test array are projected onto a plane higher than the meter.

An unbalanced data set classification problem trains two neural networks whose error functions are set as E respectively₊₁ and E_–1 defined a

$\begin{matrix} E_{+ 1} = \frac{1}{2 M_{B}} \sum_{i | y = 1} (y_{i_{i}} - o_{i})^{2} \\ + \frac{C_{+ 1}}{2 M_{A}} \sum_{i | y = 1} {(w_{+ 1}^{T}}_{i} φ_{+ 1} (x^{i}) + b_{b + 1})^{2} \end{matrix}$ (4) $\begin{matrix} E_{- 1} = \frac{1}{2 M_{A}} \sum_{i | y = 1} (y_{i_{i}} - o_{i})^{2} \\ + \frac{C_{- 1}}{2 M_{B}} \sum_{i | y = - 1} {(w_{- 1}^{T}}_{i} φ_{- 1} (x^{i}) + b_{b - 1})^{2} \end{matrix}$ (5)

Formula: y_i∈–1,+1 is the class label of the lass_i = +1 A corresponding b to find the partial derivative, and the rule of minimizing error is obtained. Error bias to weight w is:

$\frac{\partial E_{+ 1}}{\partial W_{+ 1}} = \frac{\partial}{\partial W_{+ 1}} [\begin{matrix} (\frac{1}{2 MB} \sum_{i | ji = - 1} (y_{i} - o_{i})^{2} \\ + \frac{C_{+ 1}}{2 M_{A}} \sum_{i | y_{i} = 1} {(w_{+ 1}^{T}}_{i} φ_{+ 1} (x^{i}) + b_{b + 1})^{2} \end{matrix}]$ (6)

Formula (6) item 1 is set as E_MSE +1, subparagraph 2 is set as E; and_twin +1E, I_MSE +1 For o_i = f. Net_i), o, etc_iFor the output of the neural network, f(·) is the activation function. The following relationships are available: $\frac{\partial E_{+ 1 MSE}}{\partial b_{+ 1}} = \frac{\partial E_{+ 1 MSE}}{\partial O_{i}} \times \frac{\partial o_{i}}{\partial {net}_{i}} \times \frac{\partial {net}_{i}}{\partial w_{+ 1}}$ (7)

$\frac{\partial E_{+ 1 MSE}}{\partial W_{+ 1}} = - \frac{1}{M_{B}} \sum_{i | yi = - 1} (y_{i} - o_{i}) \times f^{'} ({net}_{i}) \times φ_{+ 1} (x^{i})$ (8)

E_MSE +1b Yes₊₁The partial derivative is: $\frac{\partial E_{+ 1 MSE}}{\partial b_{+ 1}} = \frac{\partial E_{+ 1 MSE}}{\partial O_{i}} \times \frac{\partial o_{i}}{\partial {net}_{i}} \times \frac{\partial {net}_{i}}{\partial b_{+ 1}}$ (9) $\frac{\partial E_{+ 1 MSE}}{\partial b_{+ 1}} = - \frac{1}{M_{B}} \sum_{i | yi = - 1} (y_{i} - o_{i}) \times f^{'} ({net}_{i}) \times 1$ (10)

Eitem 2_twin +1A partial derag equations can be summar

$\frac{\partial E_{+ 1 twin}}{\partial w_{+ 1}} = \frac{C_{+ 1}}{M_{A}} \sum_{i | y_{i} = + 1} (w_{+ 1}^{T} φ_{+ 1} (x^{i}) + b_{+ 1}) φ_{+ 1} (x^{i})$ (11)

$\frac{\partial E_{+ 1 twin}}{\partial b_{+ 1}} = \frac{C_{+ 1}}{M_{A}} \sum_{i | y_{i} = + 1} (w_{+ 1}^{T} φ_{+ 1} (x^{i})) + b_{+ 1}$ (12)

In summary, the following equations can be summarized:

$\begin{matrix} \frac{\partial E_{+ 1}}{\partial w_{+ 1}} = - \frac{1}{M_{B}} \sum_{i | y_{i} = - 1} (y_{i} - o_{i}) \times f^{'} ({net}_{i}) \\ \times φ_{+ 1} (x^{i}) + \frac{C_{+ 1}}{M_{A}} \sum_{i | y_{i} = + 1} (w_{+ 1}^{T} φ_{+ 1} (x^{i}) + b_{+ 1}) φ_{+ 1} (x^{i}) \end{matrix}$ (13)

$\begin{matrix} \frac{\partial E + 1}{\partial b + 1} = - \frac{1}{MB} \sum_{ilyi = - 1} (yi - oi) \times f^{'} (neti) + \\ \frac{C + 1}{MA} \sum_{ili = + 1} (w_{+ 1}^{T} φ_{+ 1} (x^{i})) + b_{+ 1} \end{matrix}$ (14)

$\begin{matrix} \frac{\partial E - 1}{\partial w - 1} = - \frac{1}{MA} \sum_{ilyi = + 1} (yi - oi) \times f^{'} (neti) \times φ - 1 (x^{i}) \\ + \frac{C - 1}{MB} \sum_{ilyi = - 1} (w_{- 1}^{T} φ_{- 1} (x^{i} + b_{- 1})) φ - 1 (x^{i}) \end{matrix}$ (15)

$\begin{matrix} \frac{\partial E_{- 1}}{\partial b_{- 1}} = - \frac{1}{M_{A}} \sum_{i | y_{i} = + 1} (y_{i} - o_{i}) \times f^{'} ({net}_{i}) \\ + \frac{C_{- 1}}{M_{B}} \sum_{i | y_{i} = - 1} (w_{- 1}^{T} φ_{- 1} (x^{i})) + b_{- 1} \end{matrix}$ (16)

The steps of predicting sample x class abel are as follows: first, the x is mapped to space by hidden layer φ(· n the output er determines the x class label y: according to the following formula $y = {\begin{matrix} + 1 \frac{W_{+ 1}^{T} φ (x) + b_{+ 1}}{∥ w_{+ 1} ∥} ⩽ \frac{w_{- 1}^{T} φ (x) + b_{- 1}}{∥ w_{- 1} ∥} \\ - 1 else \end{matrix}$ (17)

Exte g the neural mirror network to the largest data set is called y_i∈ the rror network is implemented by training an independent K neural network. each network is cost-effectively determined. if the sample falls into this category, Fig. 2 shows the overall structure of the system, where each category first involves a number of giants and then a number of neuronal subclasses, where the output label of A class of neurons is z. a₁^Az, I₂^Az, z.3_p^A. The number of neurons may vary from category to category, and each output neuron of a category is hyperspectral, predicting whether new samples fall into this category. Suppose a sample of a A class, TWN will have to look for or be closer to A class, and the U must be more than 1 apart from the other categories. Structure of Twin Neural Network as show in Fig. 3. General Structure of Multi-Twin Neural Networks as show in Fig. 4.

Fig. 3

Structure of Twin Neural Network.

Fig. 4

General Structure of Multi-Twin Neural Networks.

Because the loss function of the dual neural network does not contain hyperparameters, the multi-level neural network does not need to adjust the hyperparameters of the model, but only predicts the number of hidden neural networks connected to the network. h is the error transmitted to the first feature map₁ h are the errors transmitted to the additional graph₂ h; and₃ h; and₄ h; and₅The element of the error matrix is set to M×N. E, size of the graph The error of the relevant filter is calculated as: $\begin{matrix} E_{i, j} = h_{1, i, j} - h_{2, i, j - 1} + h_{2, i, j} \\ - h_{3, i - 1, j} + h_{3, i, j} - h_{4, i - 1, j - 1} + \\ h_{4, i, j} - h_{5, i - 1, j} + h_{5, i, j - 1} \end{matrix}$ (18)

Type: M > i > 1, NM > i > jM > i > 1.

The errors transmitted to the neurons at the top of the feature map are:

$E_{i, j} = {\begin{matrix} h_{1, i, j} + h_{2} + h_{3, i, j} + h_{4, i, j} i = 1, j - 1 \\ h_{1, i, j} - h_{2, i, j - 1} + h_{3, i, j} + h_{5, i, j + 1} i = 1, j = N \\ \begin{matrix} h_{1, i, j} + h_{2, i, j} - h_{3, i, j} - h_{5, i - 1, j} i = M, j = 1 \\ h_{1, i, j} - h_{2, i, j - 1} - h_{3, i - 1, j} + h_{4, i - 1, j - 1} i = M, j = N \end{matrix} \end{matrix}$ (19)

The er ons of the fe e map is:

$E_{i, j} = {\begin{matrix} h_{1, i, j} - h_{2, i, j - 1} + h_{2, i, j} + h_{3, i, j} + h_{4, i, j} + h_{5, i, j - 1} \\ i = 1, 1 < j < N \\ h_{1, i, j} - h_{2, i, j - 1} + h_{2, i, j} - h_{3, i - 1, j} - h_{4, i - 1, j - 1} - h_{5, i - 1, j} \\ i = M, 1 < j < N \\ h_{1, i, j} + h_{2, i, j} + h_{3, i, j} - h_{3, i - 1, j} + \underset{4, i, j}{h} - h_{5, i - 1, j} \\ 1 < i < M, j = 1 \\ h_{1, i, j} - h_{2, i, j - 1} + h_{3, i, j} - h_{3, i - 1, j} - h_{4, i - 1, j} - h_{5, i, j - 1} \\ 1 < i < M, j = N \end{matrix}$ (20)

5 Financ l performance evaluation of intel nt manufacturing enterprises

This study uses a fuzz ural network method to analyze the results of company pe nce manag and multiply the previous data to get the results. Ideal 5. The calculation results show that the enterprise indicators are as follows:

A members ctor A11 the company’s profitability

$\begin{matrix} Q 11 = (0.3986, \dots 0.2340, \dots 0.1504, \dots 0.0875) * \\ [\begin{matrix} 00.870.0400 \\ 0.770.26000 \\ 0.260.430.7500 \\ 0.260.74000 \\ 00.670.4700 \end{matrix}] \\ = (0.2381, \dots 0.6458, \dots 0.136, \dots 0.0000, \dots 0.0000) \end{matrix}$ (21)

The rest o the rele nt data e calcul ed in the same way, so it is not described too much in this paper. Attribution Vector A1 Realistic Financial Perform ce Indic ors

$\begin{matrix} Q 2 = (0.4454, \dots 0.2638, \dots 0.1429) * \\ [\begin{matrix} 0.23810.64580.136000 \\ 0.61310.05420.31390.03580 \\ 0.07850.643500.0250.2603 \end{matrix}] \\ = (0.2625, \dots 0.4637, \dots 0.2138, \dots 0.0429, \dots 0.0391) \end{matrix}$ (22)

A embers p vector 2 the en rise’s potential development capability index $\begin{matrix} Q 2 = (0.1643, \dots 0.5369, \dots 0.2960) * \\ [\begin{matrix} 0.04780.23280.49810.19740.0638 \\ 0.15320.48430.31090.06260 \\ 0.18210.16240.54830.08360 \end{matrix}] \\ = (0.1453, \dots 0.3512, \dots 0.4067, \dots 0.0878, \dots 0.0121) \end{matrix}$ (23) $\begin{matrix} Q = (0.75, \dots 0.25) \\ * [\begin{matrix} 0.27520.46370.21380.04190.0391 \\ 0.14530.32510.40470.08770.0201 \end{matrix}] \\ = (0.2433, \dots 0.4145, \dots 0.2630, \dots 0.0569, \dots 0.0341) \end{matrix}$ (24)

By or ally co ning t above c lation results with the comment set, the financial performance re lts of t company an be obtain a ows:

Evaluation f profit ility A1 results $\begin{matrix} S 11 = U^{*} Q = 95 * 0.2381 + 82.5 * 0.6458 \\ + 67.5 * 0.1360 + 50.0 * 0.0000 \\ + 20 * 0.0000 = 85.078 (mark) \end{matrix}$ (25)

The other related data calculation results are the same, in this pap no lo abor ion. As shown in Table 8, the following is the final evaluation of the company’s financial performance:

Table 8

S Corporate financial performance evaluation results

Target layer	Value	Level I indicators	Value	Secondary indicators	Value
S A of Financial Performance Evaluation	78.9409	A1 of realistic financial performance	78.5092	A11 profitability	85.0780
				A12 solvency	82.9337
				Capacity A14 for development	72.5685
				A13 of operational capacity	73.4569
		A2 of potential development capacity	79.3727	A21 of external market capacity	80.1346
				A22 of internal management capacity	85.1123
				A23 of innovative capacity	72.8712

According to the analysis, S company has 78.90 points of financial performance evaluation, which can reflect its good financial performance at present. As can be seen from the results, the relatively poor performance of S company’s financial performance at present is that of development ability (72.5664 points), innovation ability (72.8453 points) and operation ability (73.4523 points). The above evaluation, we can draw a conclusion, relying on the calculation of secondary indicators can be clear, enterprises in the development capacity of the three indicators there are great differences.

5.1 Capacity development results analysis

Analysis of the development ability of the enterprise shows that the total assets of the S company are relatively weak and the growth of main business activities is slow, which indicates that the profit margin of the company is low and even negative growth trend has appeared in the past two years.

5.2 Innovation capacity outcome analysis

With the further development of our socialist market economy, the traditional financial performance management methods no longer meet the needs of modern enterprises and current economic markets. Enterprises should not look for business problems in the old financial system. They should focus on innovation and scientific research, taking due account of the sustainability of enterprise development. Generally speaking, there is an obvious positive correlation between R & D ability and enterprise competitiveness, that is, the higher the scientific research level of the company, the greater the competitiveness of the company, but it is believed that S company’s innovation ability is relatively weak.

5.3 Operational capacity outcome analysis

The analysis of the business ability of the enterprise shows that the business ability of the enterprise is very limited, the flow rate of accounts receivable and the overall liquidity ratio of the enterprise are relatively high. Low data show that the company’s liquidity ratio and working capital ratio play an important role in the development of enterprises.

6 Conclusion

This document uses fuzzy neural networks and dual data models to analyze indicators such as market performance, management level, innovation ability, solvency, enterprise development and profit margin of intelligent manufacturing enterprises to determine financial indicators, and to analyze these indicators, which accurately reflect the company’s specific situation, scientifically design and use the weight of these indicators of non-financial entity decisions. The discrete matrix model identifies two main problems faced by S companies: this paper holds that in order for S companies to achieve rapid development, they must focus on evaluating financial performance. A scientific financial performance evaluation system must be based on S company’s development strategy. Analysis S the actual situation of the company’s current development, which leads to a reasonable approach to performance management, analysis of the challenges faced by the company in the performance management process, and put forward corresponding suggestions and improvement measures according to the shortcomings, in addition, it also makes the performance appraisal system more perfect. The evaluation indicators developed in this document are more feasible, scientific, in line with the reality of enterprise development, and incorporate diversified indicators such as innovation performance into the enterprise evaluation system, although they are not financial indicators. They can have a significant impact on the performance and development of enterprises.

References

Azmi

, Mohamad

and Shah

M.E.

, Ethical investments and financial performance: An international evidence, Pacific-Basin Finance Journal 45(10) (2019), 78–89.

Ojeka

S.A.

, Adegboye

, Alabi

, Afolabi

and Iyoha

, Chief financial officer roles and enterprise risk management: An empirical based study, Heliyon 5(6) (2019), 88–92.

Du Pont. Make the Door Open: Groundbreaking African American Teachers of Singing, Voice and Speech Review 13(1) (2019), 17–26.

Malmi

and Ikäheimo

, Value Based Management practices-some evidence from the field, Management Accounting Research 14(3) (2003), 304–310.

Yaghoobi

and Haddadi

, Organizational performance measurement by a framework integrating BSCand AHP, International Journal of Productivity & Performance Management 65(7) (2016), 959–976.

Dufera-Meta

, Financial Performance EVAluation, Journal of Emerging Market Finance 11(1) (2012), 1–36.

Lager

, Samuelsson

and Storm

, Modelling company generic production capabilities in process industries, International Journal of Operations & Production Management 37(2) (2017), 78–85.

Hannabuss

, Review: Sensing Sacred, Journal of Mental Health 28(1) (2019), 56–62.

Vintilă

G.P.

, Study of Effective Corporate Tax Rate and Its Influential Factors: Empirical Evidence from Emerging European Markets, Emerging Markets Finance and Trade 54(3) (2018), 88–94.

10.

Loots

, Strings Attached to Arts Funding: Panel Assessments of Theater Organizations through the Lens of Agency Theory, The Journal of Arts Management, Law, and Society 49(4) (2019), 45–49.

11.

and Rubin

, The paradox of power: Principal-agent problems and administrative capacity in imperial China (and other absolutist regimes), Journal of Comparative Economics 47(2) (2019), 203–211.

12.

Mojumder

J.C.

, Ong

H.C.

, Wen

T.C.

, et al. The intelligent forecasting of the performances in PV/T collectors based on soft computing method, Renewable & Sustainable Energy Reviews 72 (2017), 105–118.

13.

and Hong

, Combining re-sampling with twin support vector machine for imbalanced data classification, Inter-National Conference on Parallel & Distributed Computing 42 (2017), 46–53.

14.

Gupta

, Richhariya

and Borah

, A fuzzy twin support vector machine based on information entropy for class imbalance learning, Neural Computing & Applications 3(3) (2018), 1–12.

15.

Hassib

E.M.

, El-Desouky

A.I.

, Labib

L.M.

, et al. WOA+BRNN: An imbalanced big data classification framework using Whale optimization and deep neural network, Soft Computing 12(3) (2019), 1–20.

16.

Patil

S.S.

and Sonavane

S.P.

, Enriched over_sampling techniques for improving classification of imbalanced big data, IEEE Third International Conference on Big Data Computing Service & Applications 56 (2017), 79–82.

17.

, Cheung

Y.M.

and Tang

Y.Y.

, GOBoost: G-mean optimized boosting framework for class imbalance learning, Intelligent Control & Automation 27 (2016), 406–413.

18.

Zhong

, Feng

, Liu

, et al. Sat CNN: satellite image dataset classification using agile convolutional neural net-works, Remote Sensing Letters 8(2) (2017), 136–145.

19.

Wang

J.J.

, Pahlm

, Wagner

G.S.

, et al. Validation of the vessel-specific leads(VSLs) for acute ischemia detection on a dataset with non-ischemic ST-segment deviation, Computing in Cardiology Conference 45 (2016), 88–94.

20.

Kellndorfer

, Walker

, Pierce

, et al. Vegetation height estimation from Shuttle Radar Topography Mission and National Elevation Datasets, Remote Sensing of Environment 93(3) (2004), 339–358.

21.

Müller

, Seq State: primer design and sequence statistics for phylogenetic DNA datasets, Appl Bioinformatics 4(1) (2005), 65–69.

22.

Mohapatra

R.K.

, Majhi

and Jena

S.K.

, Classification performance analysis of MNIST Dataset utilizing a Multi-resolution Technique, International Conference on Computing 78(4) (2016), 12–19.

23.

Feng

C.M.

and Wang

R.T.

, Performance evaluation for airlines including the consideration of financial ratios, Journal of Air Transport Management 6(3) (2000), 133–142.

24.

Soltan

and Mostafa

, Lean and Agile Performance Framework for Manufacturing Enterprises, Procedia Manufacturing 2 (2015), 476–484.

25.

El-Tamimi

A.M.

, Abidi

M.H.

, Mian

S.H.

, et al. Analysis of performance measures of flexible manufacturing system, Journal of King Saud University Engineering Sciences 24(2) (2012), 115–129.

26.

Shah

, Etienne

, Siadat

, et al. (Value, Risk)-based Performance Evaluation of Manufacturing Processes, IFAC Proceedings Volumes 45(6) (2012), 1586–1591.