Prediction model of regional economic development potential based on data mining technology

Due to the differences between regions, the regional economy is also affected by the development of the real economy and the environment in a specific region. Thus, it is necessary to predict the development potential of the regional economy to ensure the smooth operation of the economy. The continuous upgrading of technology enables the country to grasp the development trend of the regional economy more accurately than ever before. However, the existing prediction techniques still have large errors. Therefore, this paper constructed a prediction model based on data mining technology to predict the development potential of the regional economy to enhance the accuracy of predictive models. The experiments showed that the prediction accuracy of the prediction model constructed in this paper could reach 96.78%, which could provide an accurate reference for the development of the regional economy to ensure the stable development of the regional economy, and has important application value.


INTRODUCTION
If the overall economy of the country is to run smoothly, it is necessary to carry out scientific and rational planning for the development of the regional economy to ensure the scientific and sustainable development of the regional economy. However, its development has a certain nonlinearity, so it is very difficult to predict. In addition, the uncoordinated development of regional economies have also caused great disadvantages to the country's overall economic development. Therefore, it is well worth predicting the development potential of the regional economy in order to formulate reasonable economic policies. 1 The existing prediction models include time series model and ant colony search model. However, the prediction accuracy of these models is still poor. Therefore, experimental simulations are continuously conducted to build models with high prediction accuracy, thereby strengthening the ability to predict regional economies.
Data mining (DM) technology is mainly aimed at the processing technology of massive data information, so it can better predict the potential of regional economic development. This article combines DM technology to carry out experiments. DM can obtain more accurate economic data to, improve the accuracy of regional economic development potential prediction, and provide a more favorable basis for the formulation of economic policies. It makes accurate predictions for the development potential of the regional economy, which can promote the balanced development of regional economies. Macro and micro policies are formulated in line with regional development, which promote the overall economic stability of the country. It can alleviate the limitations of traditional forecasting models and make the forecasting method of the development potential of the regional economy to a higher level. It makes the forecasting process of economic development potential more simplified and more accurate. 2 Regional economic development has an important impact on the overall economic development of the country. Since this is a national strategic research goal, there are many scholars who study it. Among them, Bi proposed a prediction model of regional economic development potential model, which used the traditional support vector machine (SVM) method to solve the nonlinear characteristic problem of the regional economic impact index. The results showed that the model had higher prediction accuracy compared to other prediction models. 3 However, his research lacked theoretical basis and failed to comprehensively take into account the factors affecting regional economic development. Wang analyzed the forecast model and coordinated development of regional economic development potential. He proposed that through the adjustment of the macro-policy system, a macro-policy system such as market system adjustment, regional policy adjustment, and government function transformation should be constructed. It is the development direction to enhance the potential of economic growth in the context of the new normal. 4 His research lacked robust data and failed to take into account the impact of micro-policy systems on regional economic development. In the model established by Zou, he used the least squares method to optimize the nonlinear characteristic indicators of regional economic development, and obtain the initial data for predicting regional economic development. The analysis of the experimental results showed that the model had high prediction accuracy and could accurately predict the regional economic development trend, which could provide a reference for regional development prediction. 5 The data obtained by the least squares method used in his research had large errors and the data was not accurate enough. Sava used the research method of econometric model to explore the establishment of a regional economic development forecast model. The results showed that they determined some of the effects on the functioning of the entire country's economic system. 6 His research lacked strong experimental data and failed to accurately point out how the environment affects economic affairs. There are certain deficiencies.
The prediction model of regional economic development potential constructed in this paper has the following innovations: (1) This paper introduces data mining technology into the prediction model of regional economy, which makes the original data obtained by the prediction model more accurate. The predicted crystal oscillators are also more accurate.
(2) In this paper, the particle optimization algorithm is introduced into the purchased model, which can optimize the prediction data of the model and find the optimal solution. It continuously and accurately predicts the regional economic development potential, and provides data support for the formulation of regional economic development policies. 7

Data mining technology
DM technology is an interdisciplinary subject that can provide an effective way for regional economic development decision-making and support. 8 In this paper, the main task of data mining technology is to dig out accurate data of regional economic development. It excavates the laws of regional economic development from the data samples, and makes accurate predictions for the development potential of the regional economy. 9 An overview of data mining technology is shown in Figure 1: By discovering the internal relationship of regional economic development from a large amount of data, it predicts the development potential of regional economy according to the internal relationship. 10 Of course, in order to complete the task of prediction and association analysis, data mining needs to use association rule algorithm, radial basis function (RBF) network and particle optimization algorithm. Table 1 is a comparison of various DM algorithms: Genetic algorithm is mainly used for clustering and optimization, and has efficient features. Neural network can predict and classify, but the interpretation effect is general. Association rules are mainly used for classification and clustering. Statistical analysis is generally used for clustering, and the results are relatively accurate. Support vector function is used for classification, and the error value is also small. Bayesian network is mainly used for classification, clustering and prediction.  The association rule algorithm needs to divide the data in the database into two categories, namely type A and type B. Then the data in the table is analyzed and the correlation of the data is found out. Assuming that there is data (x 1 , x 2 , x 3 , … , x i ) in class A and data (y 1 , y 2 , y 3 , … , y n ) in class B, the residual matrix in the two samples is as follows: (1) Then the association between type A data and type B data can be analyzed by the following formulas: Among them, K represents the numerical difference between the two sample data, and R represents the correlation between A-type and B-type data. It is doing better clustering of these data through RBF neural network.
The RBF neural network is a three-layer neural network. 11 The structure of the RBF neural network is shown in Figure 2: The RBF neural network obtains data from the database in the data mining technology. First, all data information are shunted into the input layer. The algorithm of the number of data shunting after the data enters the input layer is as follows: Among them, v represents the speed of the data stream and h is the time it takes for the data stream to be transmitted to the input layer. In the hidden layer, there is an activation function of the form: In the formula, d represents the input vector of the input hidden layer in the data stream, y represents the output vector of the jth hidden layer node, and represents the normalization constant of the jth hidden node. The amount of data output by the last network in t p can be expressed as: In the process of data transmission, RBF neural network causes certain data loss due to the problem of transmission speed. In order to ensure the integrity of the data, the least squares method is usually used to calculate, and its least squares loss function can be expressed as: In this way, the accuracy of the data and the accuracy of the prediction results can be guaranteed. Finally, the calculation of the total amount of data output by the RBF neural network is as follows: Among them, w is the weight in the RBF neural network, and the algorithm for the total amount of output data is as follows: In order to make the predicted data more accurate, a particle optimization algorithm is also introduced in this paper. The particle optimization algorithm can find the optimal predicted regional economic development potential measurement data by reducing the error. 12 It is to continuously approach the target within a certain area. The motion model of the particle optimization algorithm to reduce the error is generally as follows: In the formula, the coordinates of one of the particles are (s, r). The coordinates of the other particle are (g, h), and ×2 represents the error between the two particles. All particles are calculated by coordinates and then gradually approach the optimal particle, and the principle of gradually reducing the error is as follows: Among them, the third particle is found through two particles (s, r) and (g, h). The third particle is closer to the optimal particle than the other two particles, and then the fourth particle (t, v) and the third particle are introduced to continue to reduce the error, as follows: In this way, the error is continuously reduced, and the optimal particle is continuously approached. The introduction of particle optimization algorithm into the prediction of regional economic development potential can make the predicted data points more accurate, and improve the accuracy of data mining technology for regional economic development potential prediction. 13 The specific data mining process is shown in Figure 3:

Development of regional economy
Development of regional economy is affected by the political, cultural, and geographical environment in a certain region. 14 The relationship between the connotation and extension of the objective law of regional economy involves economic development, and refers to exchanges between regions in terms of commodities, labor services, capital, technology, and information. 15 At present, there are still a lot of problems in the development of regional economy. One of them is the unbalanced economic development between regions, which is caused by many reasons. 16 For example, the inclination database Select data particle optimization algorithm extract information transform data preprocess data predict data F I G U R E 3 Data mining process of national economic policies increases the distance between regions, resulting in differences in economic types and industrial structures between regions, resulting in huge differences in the level of regional economic development. 17 There is also the influence of natural environment and geographical environment factors. Therefore, the difference in natural and geographical environment also leads to the imbalance of economic development between regions. The factors affecting regional economic development are shown in Figure 4: Figure 4 shows the influencing factors. Although there are different human and geographical environments between regions, there are also favorable factors for economic development within regions. Therefore, the factors that promote economic development within the region need to be carefully analyzed to find a method suitable for regional economic development, and formulate relevant policies according to local conditions. 18 For example, China's central and western regions are rich in natural resources, and natural resources are an established condition for economic development. It is necessary to formulate relevant policies to make reasonable use of these natural resources, that is, to turn natural resources into advantages for economic development.
Of course, if regional differences in economic development are to be narrowed and economies can develop together, regional protectionism needs to be broken. 19 For example, policies such as international trade barriers and commodity monopoly have not only caused blockage of circulation channels, but also prevented the free flow of various commodities and production factors. Moreover, it makes the advantages of regions, industries and enterprises unable to complement each other, resulting in waste of resources, and ultimately making it difficult to form a competitive and open market. 20 Only by promoting regional economic integration and obtaining competition for the common and fair development of the regional economies can mutually beneficial cooperation achieve win-win results. Therefore, exchanges between regions should be strengthened to seek common development. At the same time, the country's economic policy should also favor Regional strategy and policy factors Regional economic development Population resources and environmental factors Development foundation and industrial structure level factor

Regional marketization degree and institutional innovation ability
Informatization level and technological innovation ability F I G U R E 4 Factors affecting regional economic development underdeveloped regions. Taking China's western talent introduction plan as an example, the eastern part of China has a developed economy and a relatively high educational quality, and a large number of talents accumulate in the eastern region, resulting in a surplus of talents. Therefore, the state formulates the western talent introduction plan to promote the economic development of the western region and improve the problem of unbalanced regional economic development. 21 To guide its further development, it is necessary to forecast the regional economic development to reduce the impact of uncertain factors on regional economic activities. It enables decision makers to increase their understanding of future regional economic development data and minimize uncertainty or ignorance.

Construction of regional economic development potential prediction model
The prediction of regional development potential is mainly based on the regional economic development information in previous years to predict the future economic development level, so the selection of accurate historical economic development data plays a huge role in the prediction. Reasonable data mining methods are used to select the regional economic development data of previous years from the database. Then, the selected data are deeply mined to find out the internal law of regional economic development, which is used for prediction. The regional gross domestic product (GDP) development indicators cover a wide range, as shown in Figure 5: Regional economic forecast generally uses certain data analysis techniques to analyze the data of the historical development of the regional economy and discover its inherent development laws, so as to predict the future economic development direction or development potential. The regional economic development is affected by the natural geographical environment, so the prediction model constructed in this paper needs to take into account the factors that affect the regional economic development. These factors are set as {r 1 , r 2 , r 3 , … , r i }. r i represents the ith interference factor, and G is used to represent the economic development level of the region. The relationship between G and the factors that interfere with economic development is: Therefore, the prediction model constructed in this paper needs to take these interference factors into account, so as to accurately predict the development potential of the regional economy. First, it is necessary to mine the specific data of these interference factors through data mining technology, and then bring it into the construction model. The constructed prediction model function is as follows: Regional fixed asset investment import and export of foreign trade investment in environmental protection energy consumption regional retail sales of consumer goods investment in education Regional GDP F I G U R E 5 Regional  Among them, W is the weight in the RBF neural network, and the predicted GDP value obtained is continuously reduced by the particle optimization algorithm to improve the accuracy of the model prediction. Because the influence index of the regional economy is nonlinear, the development of the regional economy also has very shallow nonlinear characteristics. Therefore, the ability of data mining technology to analyze pre-data is the most scientific and accurate. It is very helpful for analyzing nonlinear data. The building of a prediction model based on data mining can better predict the development of regional economy, and also provide a favorable reference for the formulation of regional economic policies. The model prediction process is shown in Figure 6: Figure 6 is a flow chart of prediction. Firstly, data mining technology is used to analyze the factors affecting economic development and the regional economic development data in previous years from the database. Then, it is brought into the prediction model data to predict the future economic development potential. The predicted data is brought into the particle optimization algorithm to reduce the prediction error until the optimal prediction data is found, and then the data is recorded and analyzed in a unified manner.

Experiment 1
For better experimentation, this experiment selects the total economic development of a certain region from 2015 to 2020. This experiment then uses the forecasting model to forecast the total economic development of the region, and compares the forecast data with the actual economic total to judge the accuracy of the new model forecast. The total economic development of the district from 2015 to 2020 is shown in Table 2: Then, the proposed predictive model is used to predict the economic data of 2015-2020 according to the economic data of 2010-2014, so as to calculate the accuracy of the prediction of this model. According to the economic forecast from 2015 to 2020, the value continues to increase, from 2148.692 billion yuan in 2015 to 7148.692 billion yuan in 2020. The predicted economic data is shown in Figure 7: From (a) in Figure 7, it can be seen that the predicted GDP data is not much different from the actual economic data. For example, the actual economic data in 2015 is 21486.92, the economic data predicted by the model is 21486.87, with an error of 0.5. The actual economic data in 2020 is 71486.92, and the economic data predicted by the model is 71486.93, with an error of 0.1. From the perspective of error, the error range of the prediction model is mainly between 0.1 and 0.7. After the particle optimization process, the prediction error of the prediction model is getting smaller and smaller. It can be seen that the accuracy of the model in this paper can reach 96.78%. In order to verify the superiority of the model in this experiment, this experiment uses the time series model and the ant colony search model to predict the data. Then it is compared with the data predicted by the model in this article to compare whether the model in this article (the new model) is more superior. The comparison between the predicted data of the time series model and the ant colony search model and the actual data is shown in Figure 8: From Figure 8, the predicted data of the time series model and the ant colony exploration model are relatively close to the actual data. However, the prediction accuracy of the time series model is 87.45%, and the prediction accuracy of the ant colony search model is 84.37%. Therefore, the error of the prediction model is statistically calculated, as shown in Table 3: It can be seen from Table 3 that the prediction error of the time series model is between 0.2 and 1.36. Te prediction error of the ant colony search model is between 0.23 and 2.07, and the prediction error of the model in this paper is between 0.1 and 0.7. In a comprehensive comparison, the new prediction model has the lowest error, and the predicted data is the most accurate. The model needs to consider many factors in predicting economic data, so prediction also takes time.
To this end, this experiment also carries out statistics on the efficiency of the three models, and the statistical results are shown in Figure 9:   Figure 9, the prediction of the time series model takes more than 70 min. The prediction of the ant colony search model takes 40.34 min, and the prediction time required by the new model is more than 20 min. In terms of accuracy, the prediction accuracy of the time series model is 87.45%, and the prediction accuracy of the ant colony search model is 84.37%. The prediction accuracy of this experimental model is higher than that of the other two models. The prediction accuracy rate reaches 96.78%. On the whole, the new prediction model is more superior.

Experimental summary
Through experiment 1, the model in this paper can minimize the error after particle optimization algorithm to reduce the error to 0.1 and improve the accuracy of prediction. Its accuracy can reach 98.67%. Experiment 2 compares with the time series model and the ant colony search model, and it is found that the new prediction model is more superior than the other two prediction models. Not only the prediction efficiency is higher, but the prediction accuracy is also higher.

DISCUSSION
This paper expounds the prediction and association analysis tasks in data mining technology, as well as the methods of neural network and association rules, and has a certain understanding of data mining technology. However, this article only understands the tip of the iceberg of data mining technology. There are other data analysis methods for data mining technology, and each method has its own application field, such as the medical field. Therefore, DM technology can be widely used in all aspects of human life. However, a large amount of data information cannot be analyzed manually. Therefore, DM technology can not only use its existing database to store data, but also conduct in-depth analysis of the data to find the inner connection of the data, which provides great convenience for people's data processing. In this paper, the regional economy and its development are understood. The regional economy can be divided into small areas and large areas. Globally, it can be divided into European Economic Area, North American Economic Area, etc. As far as the country is concerned, it can be divided into the central economic region and the western economic region. As for how to understand the regional economy, it depends on the situation. The regional economy discussed in this paper is the economy of a certain region in a broad sense. In addition to the influencing factors mentioned in this article, each region also has its own individual influencing factors, and there are many of these factors. Therefore, the forecasting model of this paper takes into account the factors that affect the regional economic development. In order to improve the accuracy of prediction, this paper introduces particle optimization algorithm to reduce the risk of future development of regional economy.

CONCLUSION
DM technology is a technology that includes a variety of methods and algorithms, and integrates tasks such as forecasting and clustering. It can be widely used in forecasting models, especially in predicting the potential of regional economic development. It is affected by various quotations. Through DM, the internal laws of regional economic development and the relationship between other influencing factors of development can be deeply excavated, which can provide good support for the prediction of regional economic development potential. DM technology has a high degree of accuracy for the exploration of nonlinear regional economic development problems. The prediction model of regional economic development potential based on DM technology constructed has high accuracy through experiments, which is more efficient and superior than the previous prediction models. The forecasting model studied in this experiment has high practical value in the decision-making of regional economic development. DM technology can not only be applied to the processing of regional economy, but also be widely used in other researches by virtue of its own multiple algorithms and strong data processing capabilities.