WO2023231198A1 - Comprehensive evaluation method for carbon neutrality based on sparse logarithmic principal component analysis - Google Patents

Abstract

A comprehensive evaluation method for carbon neutrality based on sparse logarithmic principal component analysis, comprising the following steps: constructing a multi-dimensional carbon neutrality indicator system; performing data preprocessing on indicators; assigning a weight to second-level indicators of dimensions by using a sparse principal component analysis method; assigning a weight to first-level indicators by using an entropy weight method; and calculating a carbon neutrality index to perform comprehensive evaluation of regional carbon neutrality. According to the present invention, on the basis of an existing carbon neutrality index indicator system, the method for assigning a weight to the indicators is improved, and the sparse principal component analysis method is introduced, such that important indicators of an indictor pool can be effectively screened out, the indictor pool can be simplified, and the problem of indictor redundancy existing in a huge indictor system can be effectively avoided. According to the present invention, a first sparse main component is taken as the weight of the second-level indicators, the weight of the first-level indicators is solved by using the entropy weight method, and the index result is more robust by means of a method for fusing the weights at two stages.

Description

A comprehensive evaluation method for carbon neutrality based on sparse logarithmic principal component analysis Technical field

The invention relates to the technical field of carbon neutrality evaluation, and in particular to a carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis.

Background technique

Climate change is a global problem facing mankind. As countries emit carbon dioxide, greenhouse gases increase sharply, posing a threat to life systems. Against this background, countries around the world are reducing greenhouse gas emissions in the form of a global agreement. In September 2020, China announced to the world at the United Nations General Assembly the goal of peaking carbon emissions by 2030 and achieving carbon neutrality by 2060. The construction of an objective, systematic, comprehensive, comprehensive and dynamic carbon neutrality evaluation system to evaluate the regional carbon neutrality development level has triggered extensive research by domestic and foreign scholars.

The construction of the existing carbon neutrality index covers various indicators of economic and social transformation such as economic development, industrial characteristics, energy structure, technological innovation, finance and taxation, environmental quality, ecological governance, policy and public opinion, etc. The indicator system is large and covers a wide range of information, which can fully reflect the level of carbon-neutral development. However, the huge indicator system also has the problem of indicator redundancy. The existing carbon neutral index construction method incorporates all indicators in the indicator pool into index modeling and cannot eliminate redundant indicators.

Existing comprehensive evaluation methods are mainly divided into subjective comprehensive evaluation methods and objective comprehensive evaluation methods. The principal component comprehensive evaluation method is the main objective comprehensive evaluation method. Most of the current mainstream principal component comprehensive evaluation methods select multiple principal components based on the variance contribution rate of the principal components, and use the loadings of principal component analysis for linear weighting. When this processing method uses load coefficients in different directions, the evaluation results vary greatly, and the problem of unbalanced development in the index system cannot be identified.

Contents of the invention

The purpose of the present invention is to provide a carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis, which can solve the problems of redundant indicators and large differences in evaluation results in the evaluation index system of the prior art.

The purpose of the present invention is achieved through the following technical solutions:

A comprehensive evaluation method for carbon neutrality based on sparse logarithmic principal component analysis, including the following steps:

Collect industry information based on carbon emissions as evaluation indicators, manually index the evaluation indicators by category, and use the manually indexed evaluation indicators to build a multi-dimensional carbon neutrality index system; the multi-dimensional carbon neutrality index system includes first-level indicators and second-level indicators. level indicators;

Perform data preprocessing on the secondary indicators;

Use sparse principal component analysis method to weight the secondary indicators;

Use the entropy weight method to weight the first-level indicators;

Calculate the carbon neutrality index, and conduct a comprehensive evaluation of regional carbon neutrality based on the carbon neutrality index.

Furthermore, the first-level indicators are used to represent analysis dimensions; the second-level indicators are used to represent analysis indicators under each analysis dimension.

Further, data preprocessing for the secondary indicators includes: indicator forwarding, indicator normalization and indicator logarithmization.

Further, the indicator is positive, specifically: based on the positive and negative impact of each secondary indicator on the carbon neutrality goal, it is judged whether the secondary indicator is higher than the preset threshold, and if it is higher, the secondary indicator is set. The indicator is a positive indicator; otherwise, the secondary indicator is set as a negative indicator.

Further, the index normalization process is as shown in the following formula:

in,

is the j-th second-level indicator value belonging to the first-level indicator of the k-th dimension in the i-th region,

is the normalized index.

Further, the logarithmic process of the indicator is as shown in the following formula:

in,

is the preprocessed result of the j-th second-level indicator belonging to the k-th dimension first-level indicator in the i-th region,

is the normalized index.

Further, the process of using sparse principal component analysis to weight the secondary indicators of each dimension includes:

remember

is the preprocessed index value of the j-th among the k-th first-level indicators in the i-th region, and the indicator matrix composed of the k-th indicator dimension is:

Each indicator of the kth indicator dimension has no normalized weight coefficient β. The solution process of the weight coefficient is as follows:

in,

λ and λ ₁ are adjustment parameters, and the choice of λ ₁ plays a decisive role in the sparsity of the weight;

is the sparse first principal component loading coefficient; α represents the unit constant vector.

Represents the unitized constant vector corresponding to the kth indicator. After normalization, the weight coefficient of each indicator in the kth indicator dimension is:

Further, the entropy weight method is used to weight the first-level indicators, specifically:

Calculate the information entropy of the j-th secondary indicator among the k-th primary indicator;

Calculate the entropy weight of the j-th secondary indicator in the k-th primary indicator using the information entropy of the secondary indicator;

The weight of the kth primary indicator is calculated using the entropy weight of the secondary indicator.

Further, the formula for calculating the weight of the kth first-level indicator is:

in

is the entropy weight of the j-th secondary indicator among the k-th primary indicator,

The calculation formula is:

in

is the information entropy of the j-th secondary indicator among the k-th primary indicator,

The calculation formula is:

in

is the preprocessed index value of the j-th second-level indicator among the k-th first-level indicators in the i-th region, and n represents the region number.

Further, the calculation of the carbon neutrality index is specifically as follows:

Among them, Carbon Index _i represents the final carbon neutrality index of the i-th region,

Represents the index value of the j-th second-level indicator among the k-th first-level indicator in the i-th region that has been forwarded and normalized but not logarithmized. v ^(k) represents the entropy of the k-th first-level indicator. The weight obtained by the power method,

Represents the weight obtained by sparse principal component weighting of the j-th second-level indicator in the k-th first-level indicator.

The present invention introduces the logarithmic principal component comprehensive evaluation method, and takes the normalized load corresponding to the first principal component as the weight coefficient for geometric weighting. Taking the first principal component ensures the consistency of the index results. Geometric weighting can effectively identify the problem of unbalanced development in the indicator system and penalize areas with unbalanced development in terms of index scores. Based on the existing carbon neutrality index index system, the present invention introduces the sparse principal component analysis method into the index construction method, which can effectively screen out important indicators of the indicator pool and streamline the indicator pool.

Description of the drawings

Figure 1 is a step diagram of the carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis of the present invention.

Detailed ways

The embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings.

The following describes the embodiments of the present disclosure through specific examples. Those skilled in the art can easily understand other advantages and effects of the present disclosure from the content disclosed in this specification. Obviously, the described embodiments are only some, but not all, of the embodiments of the present disclosure. The present disclosure can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the disclosure. It should be noted that, as long as there is no conflict, the following embodiments and the features in the embodiments can be combined with each other. Based on the embodiments in this disclosure, all other embodiments obtained by those of ordinary skill in the art without making creative efforts fall within the scope of protection of this disclosure.

The carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis of the present invention includes the following steps:

Step S1: Construct a multi-dimensional carbon neutrality index system.

The multi-dimensional carbon neutrality indicator system includes two parts: primary indicators and secondary indicators.

Among them, the first-level indicators are used to represent the analysis dimensions, including at least: economic development, industrial characteristics, green industry, energy structure, infrastructure level, technological innovation, finance and taxation, ecological construction, environmental quality and green life, but are not limited to these.

Secondary indicators are used to represent specific analysis indicators under each analysis dimension, including at least:

(1) Economic development: per capita regional GDP, total import and export volume, fixed asset investment growth rate, per capita disposable income of residents, urban registered unemployment rate, urbanization rate, and population aging rate;

(2) Industry characteristics: industrial upgrading, industrial structure deviation, number of IPV4 addresses, software business revenue, e-commerce sales, sales revenue of new high-tech industry products, energy industry investment growth, petroleum processing investment growth, coal mining Industry investment growth rate;

(3) Green industry: electricity consumption per unit industrial added value, increase or decrease in energy consumption per 10,000 yuan of regional GDP, water consumption per unit industrial added value, gas emissions per unit industrial added value, wastewater discharge per unit industrial added value, investment in industrial wastewater treatment, Investment in process waste gas pollution control;

(4) Energy structure: proportion of coal consumption, proportion of crude oil consumption, proportion of natural gas consumption, proportion of electricity consumption, proportion of fossil energy power generation, proportion of power generation, and proportion of renewable energy power generation;

(5) Infrastructure level: road area in built-up areas, highway passenger volume, railway operating mileage growth rate, proportion of bus-only road length, hydropower installed capacity growth rate, and number of mobile phone base stations;

(6) Scientific and technological innovation: number of scientific and technological enterprises, number of scientific papers published, proportion of green patents, technology market turnover, and number of valid invention patents;

(7) Financial finance and taxation: fiscal revenue, fiscal expenditure, energy conservation and environmental protection expenditure, tax revenue as a share of GDP, local fiscal expenditure as a share of GDP, the number of financial practitioners, and the added value of the financial industry;

(8) Ecological construction: proportion of highway green space, green coverage rate of built-up areas, proportion of sandy soil area, newly added water and soil erosion control area, proportion of national nature reserve area, wetland area, forest area, forest coverage rate;

(9) Environmental quality: fertilizer usage, pesticide usage, carbon emissions, sulfur dioxide emissions, nitrogen oxide emissions;

(10) Green life: the harmless treatment rate of domestic waste, the number of buses per 10,000 people, the proportion of new energy vehicle consumption, and the proportion of urban green buildings in new buildings.

Step S2: Perform data preprocessing on the indicators.

Further, in the preferred embodiment of the present application, data preprocessing of indicators includes: indicator forwarding, indicator normalization and indicator logarithmization.

The positive indicator is divided into positive indicators and negative indicators based on the positive and negative impact of each secondary indicator on the carbon neutrality goal. If the indicator value is higher than the set threshold, the indicator is set as a positive indicator; otherwise, it is set as a negative indicator.

The indicator normalization process is shown in the following formula:

in,

is the j-th second-level indicator value belonging to the first-level indicator of the k-th dimension in the i-th region,

is the normalized index.

The index logarithmization process is shown in the following formula:

in,

is the preprocessed index value of the j-th second-level indicator of the k-th first-level indicator in the i-th region,

is the normalized index.

Step S3: Use the sparse principal component analysis method to weight the secondary indicators of each dimension.

The method of using sparse principal component analysis is to apply a norm penalty to the loading coefficient based on traditional principal component analysis, so that the weight of some unimportant indicators in the principal component composition is 0. The estimation process of its load sparseness is as follows:

remember

is the preprocessed index value of the j-th among the k-th first-level indicators in the i-th region, and the indicator matrix composed of the k-th indicator dimension is:

Each indicator of the kth indicator dimension has unnormalized weight coefficient β. The solution process of the weight coefficient is as follows:

in,

λ and λ ₁ are adjustment parameters, and the choice of λ ₁ plays a decisive role in the sparsity of the weight.

is the sparse first principal component loading coefficient. α represents the unit constant vector.

Represents the unitized constant vector corresponding to the kth indicator. After normalization, the weight coefficient of each indicator in the kth indicator dimension is:

The method of sampling sparse principal component analysis can effectively screen out important indicators in the indicator pool, streamline the indicator pool, and effectively avoid the problem of indicator redundancy in a huge indicator system.

Step S4: Use the entropy weight method to weight the first-level indicators.

The specific process of using the entropy weight method to weight first-level indicators is as follows:

(1) Calculate the information entropy of the j-th secondary indicator in the k-th primary indicator:

in,

Represents the normalized weight of the j-th second-level indicator in the k-th first-level indicator of the i-th region,

is the preprocessed index value of the j-th second-level indicator among the k-th first-level indicators in the i-th region, and n represents the region number.

(2) Calculate the entropy weight of the j-th second-level indicator in the k-th first-level indicator:

(3) Calculate the weight of the kth first-level indicator: Sum the second-level indicator weights of each first-level indicator to get the first-level indicator weight:

Step S5: Calculate the carbon neutrality index and conduct a comprehensive evaluation of regional carbon neutrality.

The formula for calculating the carbon neutrality index is as follows:

Among them, Carbon Index _i represents the final carbon neutrality index of the i-th region,

Represents the index value of the j-th second-level indicator among the k-th first-level indicator in the i-th region that has been forwarded and normalized but not logarithmized. v ^(k) represents the entropy of the k-th first-level indicator. The weight obtained by the power method,

Represents the weight obtained by sparse principal component weighting of the j-th second-level indicator in the k-th first-level indicator.

A comprehensive evaluation of the degree of carbon neutrality in different regions can be made based on the calculated carbon neutrality index. The higher the index value, the better the completion of carbon neutrality.

The advantages and positive effects of the present invention are:

Based on the existing carbon neutrality index index system, the present invention improves the index weighting method and introduces the method of sparse principal component analysis, which can effectively screen out important indicators of the indicator pool, streamline the indicator pool, and effectively avoid The huge indicator system has the problem of indicator redundancy. At the same time, the present invention performs logarithmic processing on the normalized index during the index preprocessing process; and uses geometric weighting to replace traditional linear algebra weighting during the index synthesis process. This processing method enables the carbon neutrality index obtained by the present invention to effectively measure the problem of unbalanced development among indicators. This invention takes the sparse first principal component as the secondary index weight, uses the entropy weight method to solve the primary index weight, and makes the index result more robust through a two-stage weight fusion method.

In the present invention, unless otherwise clearly stated and limited, the terms "installation", "connection", "connection", "fixing" and other terms should be understood in a broad sense. For example, it can be a fixed connection or a detachable connection. , or integrated; it can be mechanically connected, electrically connected or communicable with each other; it can be directly connected or indirectly connected through an intermediate medium; it can be the internal connection of two elements or the interaction between two elements, Unless otherwise expressly limited. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

The above is only for illustrating the embodiments of the present invention and is not intended to limit the present invention. For those skilled in the art, any modifications, equivalent substitutions, and improvements within the spirit and principles of the present invention can be made without creative efforts. etc., should all be included in the protection scope of the present invention.

Claims (10)

1. A carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis, which is characterized by including the following steps:

   Collect industry information based on carbon emissions as evaluation indicators, manually index the evaluation indicators by category, and use the manually indexed evaluation indicators to build a multi-dimensional carbon neutrality index system; the multi-dimensional carbon neutrality index system includes first-level indicators and second-level indicators. level indicators;

   Perform data preprocessing on the secondary indicators;

   Use sparse principal component analysis method to weight the secondary indicators;

   Use the entropy weight method to weight the first-level indicators;

   Calculate the carbon neutrality index, and conduct a comprehensive evaluation of regional carbon neutrality based on the carbon neutrality index.
2. The carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis according to claim 1, characterized in that the first-level index is used to represent the analysis dimension; the second-level index is used to represent each analysis dimension. specific analysis indicators.
3. The carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis according to claim 1, characterized in that data preprocessing of the secondary indicators includes: indicator forwarding, indicator normalization and indicator pairing. Digitization.
4. The carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis according to claim 3, characterized in that the indicators are positive, specifically: according to the positive and negative impacts of each secondary indicator on the carbon neutrality target , determine whether the secondary indicator is higher than the preset threshold. If it is higher, the secondary indicator is set as a positive indicator; otherwise, the secondary indicator is set as a negative indicator.
5. The carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis according to claim 3, characterized in that the index normalization process is as follows:

   

   in,

   

   is the j-th second-level indicator value belonging to the first-level indicator of the k-th dimension in the i-th region,

   

   is the normalized index.
6. The carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis according to claim 3, characterized in that the logarithmic process of the index is as follows:

   

   in,

   

   is the preprocessed result of the j-th second-level indicator belonging to the k-th dimension first-level indicator in the i-th region,

   

   is the normalized index.
7. The carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis according to claim 1, characterized in that the process of using the sparse principal component analysis method to weight the secondary indicators of each dimension includes:

   remember

   

   is the preprocessed index value of the j-th second-level indicator among the k-th first-level indicators in the i-th region, and the indicator matrix composed of the k-th indicator dimension is

   

   Each indicator of the kth indicator dimension has no normalized weight coefficient β. The solution process of the weight coefficient is as follows:

   

   stα ^T α＝1

   in,

   

   λ and λ ₁ are adjustment parameters, and the choice of λ ₁ plays a decisive role in the sparsity of the weight;

   

   is the sparse first principal component loading coefficient; α represents the unit constant vector,

   

   The unit constant vector corresponding to the kth indicator can be obtained after normalization. The weight coefficient of each indicator in the kth indicator dimension is:

   
8. The carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis according to claim 1, characterized in that the entropy weight method is used to weight the first-level indicators, specifically:

   Calculate the information entropy of the j-th secondary indicator among the k-th primary indicator;

   Calculate the entropy weight of the j-th secondary indicator in the k-th primary indicator using the information entropy of the secondary indicator;

   The weight of the kth primary indicator is calculated using the entropy weight of the secondary indicator.
9. The carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis according to claim 8, characterized in that the formula for calculating the weight of the kth first-level indicator is:

   

   in

   

   is the entropy weight of the j-th secondary indicator among the k-th primary indicator,

   

   The calculation formula is:

   

   in

   

   is the information entropy of the j-th secondary indicator among the k-th primary indicator,

   

   The calculation formula is:

   

   in

   

   

   is the preprocessed index value of the j-th second-level indicator among the k-th first-level indicators in the i-th region, and n represents the region number.
10. The carbon neutrality comprehensive evaluation method based on sparse logarithmic principal component analysis according to claim 1, characterized in that the calculated carbon neutrality index is specifically:

    

    Among them, Carbon Index _i represents the final carbon neutrality index of the i-th region,

    

    Represents the index value of the j-th second-level indicator among the k-th first-level indicator in the i-th region that has been forwarded and normalized but not logarithmized. v ^(k) represents the entropy of the k-th first-level indicator. The weight obtained by the power method,

    

    Represents the weight obtained by sparse principal component weighting of the j-th second-level indicator in the k-th first-level indicator.

Images