Construction of Prediction Model of Ecological Forestry Economic Development Trend Based on Deep Belief Network.

With the rapid development of economy, environmental problems become more and more serious, and people begin to pay attention to ecological construction. Forestry is an important industry related to the long-term survival and development of mankind. It undertakes the important tasks of forest product supply and ecological construction. Therefore, the quality of its development has a direct impact on all aspects of ecological economic construction. It is related to China's ecological construction and sustainable development and plays an important role in social development. The corresponding evaluation index system and evaluation model were established to deeply analyze the economic development trend of ecological forestry system. It has certain guiding significance to provide theoretical guidance for the future development direction of ecological forestry.

1. Introduction

As the main body of the terrestrial ecosystem, forest is an important prerequisite and foundation for the construction of ecological civilization. The modern forestry with the biological function of the forest as the core is more and more fundamental, subjective, and leading in the construction of ecological civilization. Since the foundation of forestry is forest, its public welfare, industrial, and weak nature are inseparable, and operation, management, and externality are all accompanied by high comprehensive risks, especially the coordinated development and continuous protection of public policies [1, 2]. At present, the implementation of ecological construction with forestry as the main body and the development of forestry ecological economy supported by forestry are becoming the construction of ecological civilization and beautiful China from both theoretical and practical levels [3]. Promoting structural reform on the supply side of the economy and realizing the Chinese dream of national rejuvenation are important contents and basic aspects.

The development of forestry ecological economy and the improvement of forestry public policy system have increasingly become national strategies, and have increasingly become the focus of academic research and social discussion [4, 5]. On the basis of promoting forestry ecological economy, the interests of relevant groups are adjusted through the implementation of policies, and the interest incentive mechanism is used to promote forest ecological construction and forestry economic growth, so as to realize the coordinated development of economy and ecology [6, 7]. As a way of economic incentive, forestry public policy not only reflects the economy of ecological protection, but also takes into account the ecology of economic development. Its implementation has theoretical basis and practical necessity.

In recent years, deep learning has developed rapidly, showing the most advanced performance in the fields of image recognition, target detection, case segmentation, and fatigue driving, and has been widely used in face recognition, pedestrian detection, automatic driving, medical imaging, industrial detection, robot vision, intelligent video monitoring, remote sensing images, and other tasks. Nowadays, the image acquisition technology has been relatively mature, and the computing power of graphics processing unit chip is also more powerful. The use of deep learning technology has become one of the research hotspots in the field of agricultural and forestry engineering. Therefore, based on the basic theory and method of in-depth learning, this article forecasts and analyzes the development trend of ecological forestry economy and provides reference suggestions for the future development of forestry ecological economy.

2. Relevant Theories of In-Depth Learning

2.1. Deep Learning

2.1.1. Overview of Deep Learning

The rise of depth learning is marked by two papers published by Hinton and his collaborators in the journals of neural computing and Science in 2006. There is a similar hierarchical structure between deep learning and traditional neural networks; that is, they all have input layer, hidden layer, and output layer. The difference is that the traditional neural network has only one or two hidden layers, and its ability to express complex functions and learn is limited. Deep learning has two to ten hidden layers, or even more. The initial value of the traditional neural network model is set randomly. Through the gradient descent method, the difference between the model output and the actual value is calculated, and the parameters are adjusted according to the difference, so as to predict the sample output. However, errors often appear “divergence” in the backward propagation; hence, the model cannot be adjusted to the best model. The deep learning adopts unsupervised layer-by-layer training, trains one layer of hidden nodes each time, and takes the output of the hidden nodes of the previous layer as the input [8]. The output of the hidden layer node of this layer is used as the input of the lower layer node. After the training is completed, the whole network is fine-tuned. It makes the performance of deep learning in most applications surpass that of other methods.

Nowadays, deep learning mostly refers to using deep neural network to train models. Deep neural network is a continuation of neural network, and neural network was one of machine learning models in the early days. Most machine learning models are developed from the field of mathematics, while neural networks are developed from biological neural networks.

The essence of deep learning is to learn the features of data and extract the high-quality representation of input data through a multilayer network composed of neurons. It makes the data linearly separable in the high-dimensional level, so that the classification and regression tasks can be carried out through a simple classifier. Deep learning uses low-level features to form more abstract high-level features. It can well simulate complex data relationships, represent data hierarchically, save useful information, and reduce the amount of data to be processed.

In research related to deep learning, researchers are more interested in improving the “depth” of the network. Deep models are made more complex by increasing the number of network layers. In subsequent studies, there will be a bias towards increasing the number of neurons and changing the way neurons are processed. Eventually, various types of deep neural networks are formed for deep learning.

2.1.2. Deep Learning Framework

When conducting research work related to deep learning, there are many frameworks to choose from, and the more popular ones are Keras, PyTorch, Caffe, TensorFlow, etc. The deep learning framework mostly adopts the multicore computing method, which has strong expressive ability and portability. A suitable framework is crucial for conducting object detection research. Through analysis and comparison, Keras framework is finally used in the training of forestry pest detection and recognition model [9].

Keras is a Python-based deep learning library, which is different from other deep learning frameworks. Keras is a high-level API specification that provides a simple and modular API to create and train neural networks, hiding most of the complex details, and can run on TensorFlow, Theano, or CNTK. It is used not only as a user interface, but also to extend the functionality of the back-end of other deep learning frameworks in which it runs. Keras started out as a simplified front-end for Theano and has since evolved to become part of Google TensorFlow, supporting both Apache MXNet and Deeplearning4j. Thanks to its widespread support, Keras' status as a tool for interframework migration is unshakable. Developers can port not only deep learning neural network algorithms and models, but also pretrained networks and weights. The Keras framework mainly has the following advantages:

Users will have a better experience with deep learning applications. The Keras API has been carefully designed to be object-oriented, flexible, and user-friendly. Researchers can define new deep learning models without having to use a very complex back-end, so the code is more concise and clear.

Seamless Python integration. Keras is a native Python package and thus provides easy access to the entire Python data science ecosystem.

Large portable working subjects and powerful knowledge base. At present, researchers mostly use Keras in combination with Theano back-end, which produces a large body of work and a strong knowledge base. It can help deep learning researchers easily back-end porting, and pretrained models can easily switch back-ends with just a few tweaks.

Keras provides a large number of learning resources and code samples for free.

2.2. Deep Learning Common Models

Common deep learning models include restricted Boltzmann machine (RBM), deep Boltzmann machine (DBM), deep belief network (DBN), convolutional neural network (CNN), and recurrent neural network (RNN).

2.2.1. Boltzmann Machine

Boltzmann machine is based on the Hopfield network by Hiton and Sejnowski, adding a random mechanism to form a random Hopfield network model, which is a probability-based model [10]. Since this model obeys the Boltzmann distribution, it is named Boltzmann machine. Boltzmann machines are symmetrical in structure and have no feedback. Its neurons consist of two parts: the hidden layer and the visible layer. The neurons in the hidden layer are called hidden nodes, and the neurons in the visible layer are called visible nodes. The Boltzmann machine neuron is a binary network. Whether it is a visible node or a hidden node, it only takes two states of 0 and 1. State 1 indicates that the neuron is activated. State 0 indicates neuron inhibition. The structure diagram of Boltzmann machine is shown in Figure 1.

Figure 1

Boltzmann machine structure.

The Boltzmann machine adopts the simulated annealing algorithm in the learning stage, which overcomes the problem that the algorithm is easy to fall into the local extreme point, and improves the ability of the algorithm to search for the global optimum point.

2.2.2. Restricted Boltzmann Machine

The restricted Boltzmann machine is proposed by Smolensky on the basis of Boltzmann machine. He restricted the connection between the original layers of the Boltzmann machine; that is, there is no connection between the same network layers. Only the nodes between the upper layer and the lower layer can be connected, which makes its probability distribution more uniform. The structure diagram of the restricted Boltzmann machine is shown in Figure 2.

Figure 2

Restricted Boltzmann machine structure.

The Boltzmann machine structure diagram shown in Figure 2 has n visible nodes and m hidden nodes. Each visible node is only related to m hidden nodes, and other visible nodes are independent of each other; that is, the state of this node is only affected by m hidden nodes. For each hidden node, it is also affected by only n visible nodes, which makes the model of restricted Boltzmann machine simple.

2.2.3. Deep Belief Network

Deep belief networks are one of the most widely used models in deep learning. Its bottom layer is composed of multiple restricted Boltzmann machines, each of which is trained separately. The top layer is a BP neural network [3, 11]. The bottom Boltzmann machine is trained by unsupervised learning, and the top BP neural network is trained by supervised learning.

Deep belief network adopts the method of layer-by-layer training; that is, one layer of restricted Boltzmann machine is trained separately each time. This process is the same as the training process of the restricted Boltzmann machine. After training the layer, the output of the layer is used as the input of another restricted Boltzmann machine and so on until every restricted Boltzmann machine has been trained. This is the pretraining process of the deep belief network. After the pretraining is completed, the top-level BP neural network is used for fine-tuning. The specific steps are as follows:

Pretraining. First fully train the first restricted Boltzmann machine, fixing its weights and offsets. Then according to the steps of the Boltzmann machine, its hidden neuron state is obtained. This state is then used as the input vector for the second Boltzmann machine. This continues until each RBM is trained, and the weights of each layer after training are retained.

Fine-Tuning. The BP neural network is set on the top layer of the deep belief network, and the BP neural network can be based on the given sample value. The supervision of the network can be properly fine-tuned. The training process of the restricted Boltzmann machine network model can be regarded as the optimization process of the parameters of a multihidden layer BP neural network. Deep belief network overcomes the problems that BP neural network falls into local extreme points due to random initialization of weight parameters and the training time is too long.

2.2.4. Convolutional Neural Network

In the 1960s, when Hube and Wiesel studied cat visual cortex cells, they found that their neurons have a unique network structure and then proposed the concept of receptive field. In the 1980s, Fukushima proposed the neurocognitive machine model based on the inspiration of the wild theory [11]. Neurocognitive machines are also considered early convolutional neural networks. In the 1990s, Le Cun et al. combined downsampling layers and convolutional layers to form the main structure of modern convolutional neural networks.

Convolutional neural network can be regarded as a special feedforward neural network model. It has the characteristics of local connection and weight sharing, and it consists of a feature extraction layer and a feature mapping layer. The structure diagram of the convolutional neural network is shown in Figure 3.

Figure 3

Convolutional neural network structure.

The feature extraction layer mainly completes the required feature extraction and recognition through local receptive fields, shared weights, and downsampling. C in Figure 3 is the feature extraction layer. Each neuron input is connected to the local receptive field of the previous layer, then passes through the next local receptive field, and then connects to another neuron in the C layer.

In Figure 3, S is the feature mapping layer, and the feature mapping layer is composed of multiple feature maps. A feature map is a plane where neurons have equal weights. The feature map structure usually uses the Sigmoid function as the activation function, so that the feature map has the property of invariant position. This property of the convolutional neural network makes the convolutional neural network have a strong advantage in image recognition after processing.

3. Ecological Forestry System

3.1. Related Concepts

3.1.1. Related Concepts

There are many discussions on the concept of forestry. The concept of “forestry” is considered in the Encyclopedia Britannica to be a business that can benefit people in the management of forestry, wasteland, and water surfaces. Its main goal is to make wood sustainable, but now the protection of water, soil, and wildlife is becoming more and more important among its goals. The concept of “forestry” in the Dictionary of Forestry Science and Technology jointly written by the International Federation of Forestry Research Organizations and the United Nations Food and Agriculture Organization is divided into three meanings. The meaning of the first layer is that forestry is a profession engaged in forest protection, construction, and management of forest land technology and business. The second level of meaning is the planting and cultivation of forest trees or the effective development and utilization of natural resources in the forest [12]. The third meaning is to maximize the benefits of human beings through the effective management and utilization of forest land and natural resources through the use of science and technology and practical experience.

On 25 June 2003, the government issued the Decision on Accelerating the Development of Forestry, which gave its own definition to the concept of forestry. Forestry is a basic industry in my country and is mainly responsible for the supply of forest products. At the same time, forestry is also a public welfare undertaking, responsible for the restoration of our country's environment and ecological construction, not only to meet the diversified needs of trees and other forest products, but also to protect the natural ecology and improve the ecological environment to meet the needs of national security. At present, the ecological demand of forestry has become the largest demand.

Forestry has the tasks of developing the economy and restoring the environment. The traditional forestry only pays attention to the economic benefits, ignoring the benefits of forestry in terms of life. Its main task is the production and development of wood, which leads to the unscientific and extensive development of forestry. When China's current forestry development direction has undergone great changes, the original development and utilization of wood based on economic resources has changed to ecological construction based on ecological resources.

3.1.2. Ecological Economy

There is no unified definition of ecological economy. The mainstream view is that ecological economy is an “economy-nature” complex ecosystem. That is, by using systems engineering methods and principles of ecological economics, all available resources can be mined within the capacity of the ecosystem [13]. Change the way of production and consumption to develop enterprises with the characteristics of ecological efficiency and economic development. Only it can have a reasonable system construction, social harmony, ecological health, and suitable landscape environment. Synthesizing the description of the concept of ecological economy by predecessors, this article believes that the ecological economy is to establish a complex ecosystem within the carrying capacity of the ecosystem. Under this system, the economy and ecology can develop in a virtuous circle. It is a mode of economic activity in which economic development and resource environmental protection are coordinated.

Ecological economy has the following characteristics. (1) Timeliness, that is, the sustainable use of natural ecological resources in time. Not only the current needs of meeting the development of the ecological economy, but also the future needs of the development of the ecological economy should be considered. (2) Spatiality, that is, each place must evaluate and recognize its strengths and weaknesses in resources rationally and scientifically, as well as scientifically and rationally utilize and develop. The government should also do a good job of resource allocation when necessary and transfer resources from rich areas to poor areas. (3) Efficiency, that is, the use of low energy consumption and high efficiency in the use of resources. The essence of ecological economy is to realize the coordinated development of ecology and economy. Taking ecological benefits as the premise and economic benefits as the basis, seek the comprehensive benefits of the combination of the two.

3.1.3. Ecological Forestry Economy

Forestry is a complex system, which involves both ecological and economic aspects. Therefore, to realize the sustainable development of forestry, we must observe the coordination and unity of these two benefits, and take the sustainable development path of forest ecology and economy interacting and coordinating with each other. In relation to forestry, human beings influence and change the forestry ecosystem through their own abilities. This impact and change is mainly manifested in two aspects: on the one hand, forestry has a significant impact on human economic activities [6, 14]. Ecological forestry provides a harmonious, stable, and beautiful natural ecological environment for the present and for future generations. On the other hand, the active action of human beings in the production activities of the forestry economy has resulted in a unique socialization of forestry.

Forestry industry includes primary forestry industry, secondary industry, and tertiary industry. The primary industry includes forest cultivation, management, pest control, seedling cultivation, forest protection, and fire prevention. The secondary industry includes forest harvesting and transportation, wood processing and comprehensive use, paper, pulp, and other industries. The tertiary industry includes commerce, catering, forest tourism, construction, transportation, forest surveying, finance, insurance, planning, and design in forestry areas.

The development of forestry has experienced the forestry industry as the top priority, and the forestry ecology is dominated, and finally, it needs to return to the economic equilibrium of forestry life. Combined with the concept definition of ecological economy described above, the forestry ecological economy defined in this article is based on the concept of ecological economy to study the development of ecological economy in forestry.

In this article, the concept of forestry ecological economy is defined as realizing the coordinated development of forestry ecology and forestry economy within the carrying capacity of forestry ecosystem. Build a complex forestry ecosystem. Under this system, the forestry economy and forestry ecology can develop in a virtuous circle, which is an economic activity mode that coordinates the development of forestry economy and resource environmental protection. Explore the regularity of coordinated development in the complex system of forestry ecology and economy, in order to study the interaction between human's forestry economic activities and the natural ecological environment of forestry and to build a sustainable development state in which human forestry economic activities and the natural ecological environment of forestry promote and coordinate with each other.

3.2. Theoretical Basis

3.2.1. Externality Theory

Externalities are also called external effects or external costs and spillover effects, and are divided into positive externalities and negative externalities according to their effect results. With regard to the definition of externality, many economists regard it as one of the most difficult concepts to define in history. But they always try to define this definition clearly, and they all give different definitions from different angles, which can be summed up from the following two angles: one is from the perspective of the subject that produces the externality and the other is to start from the perspective of the receiving subject who accepts the externality. Representatives of the first category consider externalities to be those situations in which production or consumption imposes uncompensable costs on other groups or grants uncompensated benefits. Representatives of the latter category consider externalities to be inefficiencies that arise when certain benefits or costs of an action are not considered by decision-makers.

Forestry has externalities in the production of forest products. It can have an impact on people other than its producers or consumers. This effect is divided into two cases. (1) Positive forestry externalities, that is, the benefits that the production of forest products brings to people other than the owner, and others do not need to compensate for this. (2) The negative externalization of forestry refers to the harm caused by forest products to people other than the owners. Others are not compensated accordingly. Forestry produces two types of products: one is material products (tangible products), and the other is public products (intangible products) of ecological benefits and the environment. Accordingly, forests have not only economic value, but also ecological value.

The essence of ecological forestry economy is to realize the coordinated development of ecological forestry and forestry economy. Taking the ecological forestry benefits as the premise and the forestry economic benefits as the basis, seek the comprehensive benefits of forestry by combining the two, and develop the forestry economy within the range that the ecological environment can carry. Instead of developing the forestry economy at the expense of plundering ecological forestry resources, it can achieve the healthy, stable, and sustainable development of forestry.

3.2.2. Sustainable Development Theory

The concept of sustainable development was first proposed in 1980 by the World Conservation Fund for Nature and the International Union for Conservation of Nature. It believes that sustainable development is to support economic development within the carrying range of the ecosystem to improve the quality of people's lives. The first is the sustainable development of the economy. It requires sustainable development not only in terms of economic quantity, but also in terms of economic quality. That is, while requiring sustained and stable economic growth, we must also pay attention to the protection of the ecological environment. It is required to change the economic development model to realize the three-low (low income, low consumption, and low pollution) blue production model, and to change the economic development model. Realize the transformation from extensive economic development to intensive economic development. Achieve civilized consumption and green and sustainable development. So that the economic benefits of lifting off and the protection of the environment are carried out at the same time.

Then, there is the sustainable development in ecology, which requires the development of the economy within the carrying capacity of the ecological environment. We need to obtain the primitive natural ecological resources required by human economic activities from the natural ecology. Therefore, the natural ecology is the condition and carrier for our development. We should pay attention to the protection of the ecological environment while developing the economy.

The last is sustainable development in the social aspect. The development of countries all over the world is to improve our quality of life, education, and human rights to achieve an equal and harmonious environment. No matter how different their development goals and development stages are, the purpose of their development is the same. Such purpose is the essence of sustainable development, that is, the sustainable development of society.

As we enter the 21st century, we should pay more attention to the sustainability of development. While ensuring people-oriented, the interaction and coordinated development of forestry economy and ecology are realized. On the basis of sustainable development theory, forestry ecological economy studies the coordination between forestry's life and economic system in the process of human economic activities as a whole. Make full and rational use of forestry natural ecological resources, and establish a sustainable forestry ecological economic system to achieve high-yield, high-efficiency, high-quality, and low-consumption development.

3.2.3. Coordinated Development Theory

Coordinated development theory is defined as a general term for ideas, methods, and theories that explain and solve problems of coordination and coordinated development. In 1984, biographer Ma Shihai proposed the idea of an economic-natural complex ecosystem, which was a milestone in the development of the coordinated development theory. He believes that the so-called coordinated development is to realize the mutual coexistence and coordination of ecological and economic systems in a certain area, and to form a complex system under the condition of people-oriented. The so-called coordinated development has three characteristics. The first is the development of quantity, the second is coordination in quality, and the last is dynamic continuity in time. That is to say, according to a certain number and structure, the various elements in various fields of society evolve from a low-level to a high-level process with dynamic time and space.

Forestry ecological economy emphasizes giving full play to people's subjective initiative and people's objective creativity. The constraints caused by the limited resources and environment are overcome. It realizes the growth of forestry economy under the condition of optimal utilization of forestry resources and develops the economic-ecological complex system in coordination with each other.

4. Construction and Evaluation of Economic Development Model of Ecological Forestry System

The construction of the index system is mainly the selection of indicators and the determination of the structural relationship between indicators. The establishment of the energy index system of the ecological forestry economic system is mainly based on the energy theory and set in combination with the actual situation, that is, the results of combining theory with practice. The quantitative research method is mainly used to consider the sufficiency, feasibility, stability, necessity, and coordination of the evaluation indicators. Establish the index and index structure, make the index system more scientific and reasonable and accurately and reasonably measure the development status of the ecological forestry economic system and the ability of sustainable development.

4.1. Construction of the Index System Model

4.1.1. Characteristics and Functions of the Index System

(1) Characteristics of the Index System. The energy index system of the ecological forestry economic system is faced with multilevel users, such as the public, scientific researchers, government decision-making departments, and users at different levels pay different attention to the index system. The public may pay more attention to the description and monitoring functions of the index system, while scientific researchers may pay more attention to the prediction, evaluation, and early warning functions of the index system. For government decision-making departments, it is the main responsibility of managers and decision-makers at all levels to regulate and control the sustainable development process of the ecological forestry economic system. Therefore, the energy index system of the ecological forestry economic system should be a hierarchical system [15]. The index at the lower level provides the necessary information guarantee for the higher level index, which is the basis and premise for the higher level function. The energy index system design of the ecological forestry economic system also needs to have its own characteristics, which are shown in the following points:

There is a close relationship between ecological forestry and forestry economy. It is necessary to consider the relationship between them to determine the relevant indicators. Measuring the ecological forestry economic system from the ecological and economic aspects makes the system more complete and comprehensive, and makes the energy index system more specific and convincing, which can truly reflect the current situation of the ecological forestry economic system.

The sustainability of forest ecosystem is the characteristic of sustainable development of forestry economy. The most important and sensitive indicators in the ecological forestry system should be selected in the index system. Through energy analysis, this method can well understand the ecological forestry economic system and look for a method that can make the ecological forestry economic system sustainable development, as the future development direction.

It is necessary to consider the interaction and coordination factors among subsystems in the ecoforestry economic system. Considering the subsystems in the ecological forestry economic system, so that they can develop harmoniously, people attach importance to the economy at present. The neglect of ecology leads to the decline of the self-regulation ability of the ecological forestry economic system and the poor restoration ability of the ecosystem.

The guiding ideology of the energy index system of the ecological forestry economic system is the energy theory. According to the energy theory, the basic indexes are designed, and the corresponding indexes are designed in combination with the actual situation of the ecological forestry economic system. Selecting some key indicators from the ecological forestry economic system can objectively reflect its development status and the degree of sustainable development. It is an index that focuses on the development of forestry economy and takes the index that reflects the self-status of the ecosystem as the premise.

(2) Functions of the Indicator System. The construction of an energy index system of the ecological forestry economic system is mainly based on the energy theory, in order to better analyze and discuss the internal structure of the ecological forestry economic system and the overall development and change law of the system [16]. On the basis of understanding the ecological forestry economic system, this article seeks the methods and strategies that can make the ecological forestry economic system sustainable development. Therefore, the following functions shall be ensured in the design of the index system:

Objectively reflect the status of various elements of the ecological forestry economic system. Through the establishment of the energy index system, the input and output energy of the ecological forestry economic system can be reflected. It can truly reflect the input of nature and make the input of energy more explicit.

Dynamically describe the changes of each element state of the ecological forestry economic system, and establish the energy index system of the dynamic ecological forestry economic system. The development and changes of the ecological forestry economic system in recent years are analyzed dynamically, and the development trend of the ecological forestry economic system is studied. It provides theoretical support for the sustainable development of the ecological forestry economic system.

It can infer the development trend of the ecological forestry economic system according to the status and changes of various elements of the ecological forestry economic system according to the energy theory. Understand the current development of the ecological forestry economic system in essence, and grasp its sustainable development.

According to the coordination degree among the various elements of the ecological forestry economic system, the level and ability of the sustainable development of the ecological forestry economic system are judged. At the same time, it puts forward early warning for the important problems endangering the sustainable development of the ecological forestry economic system.

4.1.2. Basic Principles and Ideas of Model Construction

(1) Basic Principles of Model Construction.

Scientific principle. Each index has a clear concept and certain scientific connotation, which can measure and reflect the current situation and development trend of the regional composite system.

Principle of practicality and operability. The indicator system is ultimately used by decision-makers to serve policy formulation and scientific management. Therefore, the principles of practicality and operability should be considered. The practical and operable index system is easy to be understood and accepted by users, and easy to collect and quantify data.

The principle of relative completeness and relative independence. As an organic whole, the index system should be able to comprehensively reflect and measure the main development characteristics and status of the assessed region [17]. In addition, the indicators describing the development of the regional composite system often have overlapping information among the indicators. Therefore, when selecting indicators, we should choose relatively independent indicators as far as possible, so as to increase the accuracy and scientificity of the evaluation.

Principle of principal component and pertinence. On the basis of completeness, the indicator system should strive to be concise and try to select those representative comprehensive indicators and main indicators. The index system should be established according to the main characteristics and existing problems of the research system.

The principle of combining inheritance with innovation. The evaluation index system of sustainable development can neither completely copy the existing index systems nor completely separate from them. Instead, it should be developed and innovated on the basis of making full use of and drawing lessons from the existing index systems of sustainable development.

(2) Thinking of Model Construction. At present, although there are few energy index systems of the ecological forestry economic system proposed at home and abroad, there are some problems in the evaluation criteria. But under the continuous exploration, the energy index system of an ecological forestry economic system will become more perfect. The basic idea of constructing energy index system of the ecological forestry economic system is as follows:

According to the energy theory and the actual situation of the ecological forestry economic system, the energy input index of the ecological forestry economic system is established. It reflects the energy input of the ecological forestry economic system from nature and economy.

Establish the index system of energy output of the ecological forestry economic system. Analyze the economic output of the ecological forestry economic system, and understand the output status and output structure of the ecological forestry economic system.

Construct the input-output index system of the ecological forestry economic system. Design indicators to better reflect the development status and sustainable development capacity of the ecological forestry economic system, and pave the way for the introduction of sustainable development indicators.

The energy index system of the ecological forestry economic system established above is summarized to synthesize the sustainable development index of the ecological forestry economic system. This index is used to measure the ecological forestry economic system, judge its sustainable development ability, and provide a theoretical basis for the suggestions.

4.2. Evaluation and Analysis on Economic Development of the Ecological Forestry System

The energy output rate is used to evaluate the output rate of the system, and the environmental load rate is used to evaluate the environmental pressure. The two represent the evaluation indicators of the sustainable development performance of the ecological forestry economic system. However, only these two indicators cannot complete the comprehensive evaluation of the system's sustainable development performance [18]. The Energy Sustainability Index (ESI) proposed by American ecologist and Italian ecologist Ulgiati. S is defined as the ratio of the system energy output rate to the environmental load rate. The evaluation standard of the ESI is 1 < ESI < 10, indicating that the system is full of vitality and development potential; when ESI > 10, it is a symbol of economic underdevelopment, and when ESI < 1, it is a consumption-oriented economic system, as shown in Table 1.

Table 1

Evaluation criteria for sustainable development capacity.

ESI	Expression meaning
ESI < 1	Consumer economy system
1 < ESI < 10	Dynamic and potential system
ESI > 10	Economically underdeveloped system

Net energy output rate reflects whether the system has advantages in obtaining economic input energy, which to some extent reflects the sustainable development of the system. Odum thinks that the value should be between one. If the net energy output rate of a system is less than 1, the energy value of the system will not increase. If the net energy output rate of one system is less than that of the other system, the opportunity and amount of economic investment for the system will be less than that of the other system. Thus, the competitiveness of the system to obtain economic input energy will be reduced, and it will eventually be replaced by the system with a high net energy output rate.

The higher the environmental load rate, the higher the technical level of the system in the utilization of energy value. On the other hand, it also reflects the greater pressure on the local environment, and a very high environmental load rate means that economic activities put too much pressure on the local environment. It may even exceed the carrying capacity of the ecological environment.

5. Conclusion

The development of ecological forestry economy has a direct impact on all aspects of ecological economic construction. It is related to my country's ecological construction and sustainable development, and occupies an important position in social development. This article introduces deep learning algorithms and their self-learning characteristics. The corresponding indicators are selected from the aspects of forestry ecology and economy, and the evaluation index system of forestry ecological economy is constructed. On the basis of predecessors' research on forestry ecological economy, it systematically analyzes the development of forestry ecological economy in my country from two aspects: horizontal and vertical. In this way, the general situation of my country's forestry ecological economic development is understood, and corresponding conclusions and policy suggestions are given.