Prediction Method of Government Economic Situation based on Big Data Analysis

The government economic situation data acquisition terminal shown in Figure 2 is responsible for signal acquisition, data processing, and wireless transmission.

As can be seen from Figure 2, the system hardware consists of a government economic situation signal detection processing board, an A/D converter, a microcontroller CPU, and a GPRS module. Through signal conditioning and A/D conversion, the substation bus voltage signal detection and processing and GPRS module control are realized. The GPRS module of the connection board realizes the data transmission among the CPU, GPRS module, and the mobile company SIM card [10]. According to the data acquisition and processing speed and performance requirements of the government's economic situation, the control chip STC12C5608AD is adopted, which has high speed, low power consumption, and strong anti-interference ability. STC12C5608AD is an enhanced 51 single-chip microcomputer with an AD function. All IO ports of the on-chip RAM768Byte can be configured. The high-level drive capability can reach 20mA. It has EEPROM function with only 20 or 28 pins, so it can reduce the volume. There is a reset circuit under the same crystal frequency; the working speed is 12 times that of the traditional 51 single-chip microcomputer. The instruction code takes into account 8,051, but its speed is 8–12 times faster than 8,051. Using this chip, the usage of A/D converter and peripheral equipment can be avoided, which greatly reduces the difficulty in circuit design and software programming [11].

The government economic situation early warning system consists of a main port and multiple branch ports. The main port is connected to multiple stations at the same time, and the branch port is connected to the main port through a network channel or serial port channel. The port consists of measurement and control device and telemetry terminal equipment. The measurement and control device are connected with the primary equipment installed on the bus through the line, collecting the government economic situation data from the primary equipment. Figure 3 shows the government's economic situation and early warning system.

It can be seen from Figure 3 that the system has the advantages of clear function, low equipment cost, convenient carrying, accurate detection, and judgment, avoids the interference of local economy and policies, and greatly shortens the elimination time of interference factors.

Combined with the cost-effectiveness of government infrastructure, the influencing factors of government economic development are extracted. Cost-benefit mainly includes the construction cost of infrastructure and operation cost. The former includes municipal, road, real estate, communication, and security facilities, which usually account for 5% to 10% of the total investment. The latter includes the consumption of labor, energy, and other materials required for the daily operation and maintenance of municipal, road, real estate and safety systems, and the maintenance costs of infrastructure, power system, heating system, water supply system, and other costs.

When analyzing the benefits of government infrastructure, it is assumed that the service life of government infrastructure is \({A }\) years, the completion time of government infrastructure and superstructure is the initial time \({t = 0}\), and the government infrastructure operator purchases infrastructure at the initial time. In time \({A }\), assuming that the income \({C}\) and annual usage \({D}\) are constant, and the project investment cost is \({E}\), including infrastructure construction cost and facility present value; under the usage \({D}\) state, the fixed cost remains unchanged, only variable cost is considered, and all costs are calculated by opportunity cost. The conditional expression of positive net present value is obtained as follows:

Where, \({F(D) }\) is the annual social income of the project, \({G(D)}\) is the annual operation and maintenance cost based on \({D}\), and \({H }\) is the investment cost, \({e}\) represents cost error. According to formula (1), during the operation of government infrastructure construction, operation and maintenance costs, labor costs, energy consumption, and so on. will be generated.

Screen the influencing factors of government economic development, build the government economic development index system, and collect the government economic index data. Combined with relevant indicators and data of economic development, 9 factors such as per capita GDP, per capita industrial production, per capita fixed asset investment, per capita fiscal income, per capita resident savings deposit balance, per capita total retail sales of social consumer goods, per capita net income of rural residents, the proportion of fiscal income in GDP and the speed of economic development are selected. This article discusses the economic development in the radiation zone of government infrastructure. The comprehensive development level of the node is characterized by the comprehensive quality index and measured by composite indicators. 14 indicators are selected to build the comprehensive development level index system from the three levels of economic development, social development, and urban construction. The maximum method is used to standardize the data, and the comprehensive development level is preliminarily measured with the help of SPSS software. The index system is shown in Table 1.

The basic data are collected from the government statistical yearbook, i.e., China Urban Statistical Yearbook 2020.

In addition to the statistical yearbook as a collection source, government economic index data can also be collected through web crawlers. Take the government's economic development as the keyword, search the relevant web pages of situation indicators on the Internet, join the crawler queue from the seed Uniform Resource Locator (URL), analyze and download the web pages, grab the URL and obtain a new URL. The data pages containing the government's economic situation are considered relevant, including all economic related pages. However, in order to simplify the calculation, only the top 10 pages of the Internet search are collected, and the same economic situation data are not collected. In the real network traffic data, count the usage heat of various indicators, read the web page on the front page of the web page, find other link addresses in the web page, find the next web page, and set the access layers of different web pages until all web pages of the web site are captured [12]. Preprocess the captured pages, take the government's economic development as the theme content, filter out the pages with inconsistent theme, use the table tag to repair and sort out the wrong or irregular tags, store the repaired complete pages in the HTML document, select the HTML file as the root node, construct the tag tree, and use the visual information of the web page to process the web page in blocks. In line with the forecast demand of the government's economic development, remove the redundant information of the web page, link the useful information together, find the text file related to the subject content, mark the hypertext, and integrate the web page [13]. Finally, through the HTTP protocol, assist the browser to download the web page, capture the effective information in the web page, including sound, text, image and other documents, obtain the government economic index data in the field of government economic development, add the government economic development related content, collect the video, audio, database, picture, text data and other types of data in the web page, eliminate new URLs, add a new crawling queue and cycle the above operations. So far, the collection of government economic development index data is completed.

The massive government economic development index data collected are preprocessed to enable the government economic index data to accurately express the government's economic situation.

Based on the semantic attention of government economic indicator data, this study highlights the semantically similar government economic development indicator data, and uses it to predict the government economic situation. According to the above basis, analyze the factors that affect the semantic distance between data items including economic development, social development, and urban construction. According to the analysis results of the above influencing factors, the influencing factors are taken as the dynamic characteristics of the government economic indicator data, so that the data of the government economic indicators change over time, showing different characteristics of government economic development. Obtain multidimensional data information according to dynamic features, conduct data exploration to reduce the dimension of government economic indicator data, and convert multidimensional dynamic features into two-dimensional dynamic features according to the dimension combination obtained by data exploration [16]. Assuming that the characteristic dimension of government economic indicator data is \({z }\), the calculation formula for the combination exploration condition \({R }\) of government economic indicator data isThe abstract features of government economic index data in information space are extracted, and the abstract feature types are divided into three categories: time series, network, and level. The calculation parameters of semantic distance are determined according to the structural relationship among the three types of government economic indicator data. Suppose that the dynamic characteristic object of the government economic development index is \({s}\) and the data object of any government economic index in information space is \({x }\), then the semantic distance \({d( {s,x} ) }\) between \({s }\) and \({x }\) is

Where, \({ f( {s,x} ) }\) is the two-dimensional display of \({x }\) on the combination of \({s }\) dynamic feature dimensions. The implicit intention is used to determine the impact of government economic index data on the prediction of government economic situation, the explicit intention is used to clarify the prediction intention of government economic situation.

\({g(s,x) }\) represents the association relationship between \({x }\) and \({s }\).

\({l(s,x) }\) is the center distance after semantic representation of \({x }\) and \({s }\), and \({w }\) represents the weight.

The semantic distance is taken as an important parameter of the semantic attention of government economic indicator data. Through \({d(s,x) }\), the distance between \({x }\) and \({s }\) at the semantic level is adopted, based on which it is possible to determine the a priori importance of different data items of government economic indicator data in the prediction of government economic situation, set the semantic attention threshold, and limit the collection of government economic indicator data items. The semantic attention \({P }\) of government economic indicator data item \({x }\) with respect to \({s }\) is

Where, \({k( {s,x} ) }\) is the a priori importance of \({x }\) with respect to \({s }\), and \({C }\) is the semantic attention threshold. The greater the semantic attention, the closer the semantics of \({x }\) and \({s }\). Aggregate the data items of government economic indicators with similar semantics to assist in the prediction of the government economic situation.

For the government economic development index data with similar semantics, the association rules of the government economic situation in the surrounding areas are mined, and the economic development model of each government is determined according to the relationship between different attributes and characteristics of the government economic index data. In the dataset, the attribute information of government economic development index data is extracted and divided into three sets: continuous attribute set, original invariant attribute set and nominal attribute set. This study uses the knowledge base of HowNet, defines the words existing in the semantic dictionary, and takes the def item in HowNet as the concept of words. According to the above concept of words, words with similar meanings need to be replaced, so that words have semantic relevance. HowNet (English name is HowNet) is a common-sense knowledge base that takes the concepts represented by Chinese and English words as the description object and regards reminding the relationship between concepts and the attributes of concepts as the basic content. At the same time, the semantic similarity interval between words is considered in this study, and the minimum semantic similarity and the maximum semantic similarity are adopted. Calculate the semantic similarity of different government economic indicator data according to the minimum and maximum semantics. The specific calculation formula is

Where, \({{K}_a,{K}_b }\) is the concept of semantic word \({a,b }\) of government economic indicator data, \({{K}_a \cap {K}_b}\) is the number of words with the same definition of the two concepts. The value of concept similarity is within [0, 1]. The smaller the similarity is, the lower the possibility of concept semantic correlation between the mined feature attributes and the prediction of government economic situation is, and the greater the similarity is, the closer the concept semantics is. Set the semantic similarity threshold, select the feature attribute with \({M }\) greater than the threshold, extract the government economic index data and determine the similarity between data. The semantic similarity matrix is used to represent the semantic similarity of all government economic index data. Combined with the semantic elements of government economic situation prediction, we mine the index features of deep semantic connection, analyze the common parts of the semantic elements of index feature attributes, and obtain the semantic connection key points and the semantic information describing the characteristics of government economic development. According to the semantic bias of economic development characteristics to economic situation prediction, semantically process the data mining results of government economic indicators, and define the characteristics of government economic development. So far, the excavation of the characteristics of government economic development is completed.

The experimental data were collected from the National Bureau of statistics. The types of data collected include residents’ income, production level, socio-economic level, and so on. Input the government economic development data into BP neural network to predict the government economic situation. The BP neural network applies the radial basis function of Multivariable Interpolation, selects the three-layer forward network as the typical structure of the neural network, and transforms the characteristic attributes of government economic development in the surrounding areas extracted from the input layer in the middle layer, so as to make the category of the characteristic attributes of government economic development in the surrounding areas closer to the center of the network. If the output value of the \({i }\)th neuron is \({{x}_i }\) and the sample point of the \({j }\)th network center is \({{G}_j }\), the corrected new network center \({B }\) isThe characteristic attributes of government economic development in surrounding areas are divided into new network centers, and the collection of network centers is used as the value domain to replace the characteristic values of government economic development in surrounding areas, so as to eliminate the impact of different dimensions of data on the prediction of government economic situation and find out the change law of government economic situation. In the prediction of the government's economic situation, with the increase of the prediction length, the error of the prediction value will become larger and larger. Therefore, the BP neural network adopts the learning training of the fitting error difference to ensure the prediction accuracy of nonlinear factors. The learning algorithm of the BP neural network is composed of four processes on the premise of the error back propagation algorithm in the neural network. The input mode in the first stage is the forward propagation of the input layer to the output layer through the middle layer, the expected output of the network and the actual output of the network in the second stage is the error inverse propagation of the error signal to the input layer through the middle layer, and the connection weight of the neural network is corrected layer by layer; the third stage is the repeated alternation of the error inverse propagation and the mode forward propagation; and the fourth stage is the convergence of the neural network, and the learning convergence process of network global error tending to a minimum [17, 18]. The overall process of predicting the government's economic situation by BP neural network is shown in Figure 4.

As shown in the three-layer BP network structure in Figure 4, the number of nodes in the input layer is set to 2, the number of nodes in the hidden layer is set to 6, the number of nodes in the output layer is set as the number of output vectors, and the number of output vectors of the target value of the neural network is set to 1, that is, the prediction result of the government's economic situation. Taking the characteristic attribute of the government's economic development as the training data and test data, BP neural network training is carried out based on training data, and the predicted value of the government's economic situation is output. The predicted value of economic situation is divided into 1 ∼ 5 levels, as shown in Table 3.

According to Table 3, the level of government economic situation is determined. So far, the prediction of the government economic situation has been completed, and the design of the government economic situation prediction method based on big data analysis has been realized.

Taking the regional economic belt around regional government infrastructure as the experimental object, the absolute economic difference is expanding year by year from 2010 to 2020, and the relative economic difference is expanding in fluctuation. It can be roughly divided into two stages, including the stage of narrowing the economic gap from 2010 to 2013 and the stage of continuous expansion of economic differences from 2013 to 2019. In the first stage, the coefficient of variation and Searle index are lower than the average level. The Searle index reached the lowest value of 0.3065 in 2012 and the coefficient of variation reached the lowest value of 0.85352 in 2013. In the second stage, the coefficient of variation and sear index in 2019 are 1.18 and 1.32 times higher than those in 2000, respectively. The experimental data comes from the economic data published by the National Bureau of statistics, and there are a total of 3,000 sets of energy consumption data, including 2,000 groups of data are randomly selected as training data and 1,000 groups of validation data. This method is used to validate the model. The number of particle swarm optimization is set to 36 and the learning factor is 2.05. 100 independent simulation experiments were carried out for each method, and 300 iterative calculations were carried out for each simulation.

The average value of the relative error between the prediction result and the actual value is regarded as the evaluation index of the model accuracy. The model simulation verification results are shown in Table 4.

Random factors may interfere with the output results in the process of data training. While calculating the training error value of each prediction model, add the average value of calculation error to avoid affecting the evaluation of the prediction model and improve the stability of output results. The comparison results of RMSE, MAE, and average error of the output results of each model are shown in Table 4.

As can be seen from Table 4, the RMSE, MAE and average error of the proposed prediction model are all less than 0.5, which has obvious advantages among the three methods. The average value of RMSE, MAE, and error of the prediction model in Reference [3] is not less than 1.61, and the average value of RMSE, MAE, and error of the prediction model in Reference [4] is not less than 1.33. It can be seen from the above comparison that the prediction model of the proposed method has a low error rate and strong prediction performance.

The proposed method, the Reference [3] method, and the Reference [4] method were compared in terms of prediction accuracy. The comparison results of prediction accuracy are shown in Figure 5.

As seen from Figure 5, the economic situation prediction accuracy of the proposed model in this article is high, which remains above 94%. The economic prediction accuracy of the Reference [3] method and Reference [4] method shows an overall deviation trend. From the above comparison, we can see that the proposed model can accurately predict the economic situation without significant deviation, and the prediction accuracy is high at any time node. It is further verified that the proposed method has high prediction accuracy and a low error rate.

Through in-depth analysis of the above prediction results, the percentage distribution of the prediction results of the proposed method, Reference [3] method, and Reference [4] method within different relative error ranges is shown in Figure 6.

It can be seen from Figure 6 that when the proposed method is used to predict the relative error, the proportion fluctuation of the error in different ranges is small, and the peak value of the overall error rate is no more than 15%, which shows that the changing trend of power supply energy efficiency predicted by the proposed method is not disturbed by external factors, and the output result is stable. The error rate of the methods in Reference [3] and Reference [4] fluctuates obviously in different ranges, and the peak error rate of the method in Reference [3] is not less than 25%. The peak error rate of the method in Reference [4] is not less than 30%. Through the above comparison, it can be seen that there is a large gap between the methods in Reference [3] and Reference [4] and the proposed methods, indicating the stability and practicability of the proposed methods.