A method and system for constructing a semantic knowledge base in the field of environmental impact assessment

Abstract

The present invention proposes a method and system for constructing a semantic knowledge base in the field of environmental impact assessment, comprising: for each industry category, constructing a professional terminology knowledge base and a non-professional terminology knowledge base corresponding to the industry category according to business needs and industry expert knowledge, and establishing a fuzzy extraction rule for each non-professional terminology keyword in the non-professional terminology knowledge base through a regular expression; constructing a four-element relationship among industry categories, professional terms, rules and keywords, and generating a keyword-rule vocabulary according to the four-element relationship; selecting corresponding professional terms from the keyword-rule vocabulary according to the industry category of a target environmental impact assessment document, locating a target chapter of the target environmental impact assessment document and a table and text content corresponding to the target chapter, extracting keywords in the target chapter and a table and text content corresponding to the target chapter according to rules in the keyword-rule vocabulary, scoring the extraction results, and evaluating the standardization of the target environmental impact assessment document according to the scoring results.

Description

Technical Field

[0001] The present invention relates to the field of environmental impact assessment, and in particular to a method and system for constructing a semantic knowledge base in the field of environmental impact assessment. Background Art

[0002] Environmental impact assessment reports are not written in a standardized manner and professional terms are not accurate. When the Environmental Impact Assessment Bureau approves a project, it determines whether the overall assessment of the project situation meets industry standards and specifications based on the environmental impact assessment report, and screens out non-compliant environmental impact assessment reports.

[0003] Every month, thousands of environmental impact assessment documents are collected by the Environmental Protection Bureau. The existing technical solution is to read, approve, and record non-compliant documents one by one manually, which is a huge workload. There are also some technical solutions in the existing technology that use information extraction technology to automatically extract document information, but environmental impact assessment reports are different from general documents. They have unified industry standards and specifications, and environmental impact assessment reports are long and contain text information and complex table information. Existing technologies are difficult to apply to the field of environmental impact assessment. Therefore, it is urgent to design a semantic knowledge base construction method and application system in the field of environmental impact assessment. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention proposes a method for constructing a semantic knowledge base in the field of environmental impact assessment, which can solve the problem of constructing an environmental assessment semantic knowledge base and the problem of normative detection and screening of environmental assessment documents, including:

[0005] Step 1: For each industry category, construct a professional terminology knowledge base corresponding to the industry category based on business needs and industry expert knowledge;

[0006] Step 2: Collect samples of environmental impact assessment documents belonging to various industry categories, locate the target chapters of the environmental impact assessment document samples through rules, extract multiple candidate keywords in the target chapters, calculate the similarity between each candidate keyword and professional terms, and collect keywords with similarity higher than the threshold to form a non-professional terminology knowledge base. For each non-professional terminology keyword in the non-professional terminology knowledge base, establish fuzzy extraction rules through regular expressions;

[0007] Step 3: Based on the professional terminology knowledge base, the non-professional terminology knowledge base and the fuzzy extraction rule, a four-element relationship among industry category, professional terminology, rule and keyword is established, and a keyword-rule vocabulary based on a tree structure is generated based on the four-element relationship;

[0008] Step 4: Obtain the target environmental impact assessment document, select corresponding professional terms from the keyword-rule vocabulary according to the industry category of the target environmental impact assessment document, locate the target chapter of the target environmental impact assessment document and the table and text content corresponding to the target chapter, extract keywords from the target chapter and the table and text content corresponding to the target chapter according to the rules in the keyword-rule vocabulary, and score the extraction results, and evaluate the standardization of the target environmental impact assessment document according to the scoring results, for example, treat the target environmental impact assessment document with a scoring result lower than a preset value as a non-compliant document.

[0009] The method for constructing a semantic knowledge base in the field of environmental impact assessment, wherein step 2 includes a rule updating step, for an environmental impact assessment document sample, the target chapter of the environmental impact assessment document sample is located through the professional terms in the professional terminology knowledge base, and the target chapter is extracted according to the fuzzy extraction rule. If the extraction result is empty, multiple candidate keywords in the target chapter are extracted, and the similarity between each candidate keyword and the professional term is calculated. The keywords with a collection similarity higher than a threshold are combined to form a non-professional terminology knowledge base and update the fuzzy extraction rule.

[0010] The method for constructing a semantic knowledge base in the field of environmental impact assessment, wherein the industry categories include livestock and poultry breeding industry, cement industry, electricity, heat production and supply industry.

[0011] The method for constructing a semantic knowledge base in the field of environmental impact assessment, wherein step 1 specifically includes: traversing the environmental impact assessment industry standard documents of various industry categories, extracting keywords from each part of the text content and all tables in the corresponding appendix, screening the keywords based on business needs and industry expert knowledge, and constructing a professional terminology knowledge base based on the screening results.

[0012] The method for constructing a semantic knowledge base in the field of environmental impact assessment, wherein the target chapter in step 2 includes a table and text; for the table, all words are extracted from the table in order as candidate keywords; for the text, the summary paragraph is extracted and segmented to obtain candidate keywords;

[0013] The fuzzy extraction rules include table form and text form, and different extraction rules are established for the table and text respectively.

[0014] The present invention also proposes a semantic knowledge base construction system in the field of environmental impact assessment, which includes:

[0015] Module 1: For each industry category, construct a professional terminology knowledge base corresponding to the industry category based on business needs and industry expert knowledge;

[0016] Module 2: Collect samples of environmental impact assessment documents belonging to various industry categories, locate the target chapters of the environmental impact assessment document samples through rules, extract multiple candidate keywords in the target chapters, calculate the similarity between each candidate keyword and professional terms, and collect keywords with similarity higher than the threshold to form a non-professional terminology knowledge base. For each non-professional terminology keyword in the non-professional terminology knowledge base, establish fuzzy extraction rules through regular expressions;

[0017] Module 3: establishing a four-element relationship among industry categories, professional terms, rules and keywords based on the professional terminology knowledge base, the non-professional terminology knowledge base and the fuzzy extraction rule, and generating a keyword-rule vocabulary based on a tree structure based on the four-element relationship;

[0018] Module 4: Obtain the target environmental impact assessment document, select corresponding professional terms from the keyword-rule vocabulary according to the industry category of the target environmental impact assessment document, locate the target chapter of the target environmental impact assessment document and the tables and text content corresponding to the target chapter, extract keywords from the target chapter and the tables and text content corresponding to the target chapter according to the rules in the keyword-rule vocabulary, score the extraction results, and evaluate the standardization of the target environmental impact assessment document according to the scoring results.

[0019] The system for constructing a semantic knowledge base in the field of environmental impact assessment, wherein module 2 includes a rule updating module, for an environmental impact assessment document sample, locates the target chapter of the environmental impact assessment document sample through the professional terms in the professional terminology knowledge base, extracts the target chapter according to the fuzzy extraction rule, and if the extraction result is empty, extracts multiple candidate keywords in the target chapter, calculates the similarity between each candidate keyword and the professional term, collects keywords with a similarity higher than a threshold, forms a non-professional terminology knowledge base and updates the fuzzy extraction rule.

[0020] The semantic knowledge base construction system in the field of environmental impact assessment, wherein the industry categories include livestock and poultry breeding industry, cement industry, electricity, heat production and supply industry.

[0021] The system for constructing a semantic knowledge base in the field of environmental impact assessment, wherein module 1 specifically includes: traversing the environmental impact assessment industry standard documents of various industry categories, extracting keywords from each part of the text content and all tables in the corresponding appendix, screening the keywords based on business needs and industry expert knowledge, and forming a professional terminology knowledge base based on the screening results.

[0022] The said system for constructing a semantic knowledge base in the field of environmental impact assessment, wherein the target chapter in the module 2 includes a table and a text; for the table, all words are extracted from the table in order as candidate keywords; for the text, the summary paragraph is extracted and the word segmentation is performed to obtain the candidate keywords;

[0023] The fuzzy extraction rules include table form and text form, and different extraction rules are established for the table and text respectively.

[0024] It can be seen from the above scheme that the advantage of the present invention is that the present invention can quickly extract and judge environmental impact assessment documents according to industry standards and specifications, and screen out non-compliant documents from a large number of environmental impact assessment reports. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a schematic diagram of the four-element relationship between industry, professional terms, rules and keywords;

[0026] Figure 2 This is a flow chart of an embodiment of the present invention;

[0027] Figure 3 and Figure 4 This is a schematic diagram of extraction results according to an embodiment of the present invention. DETAILED DESCRIPTION

[0028] Based on this, the present invention proposes a method and application system for constructing a semantic knowledge base in the field of environmental impact assessment, which quickly extracts and judges information according to industry standards and specifications, and screens out non-compliant documents from a large number of environmental impact assessment reports to solve the problems raised in the above background technology.

[0029] In order to make the above features and effects of the present invention more clearly understood, embodiments are given below and described in detail with reference to the accompanying drawings.

[0030] The present invention provides a method for constructing a semantic knowledge base in the field of environmental impact assessment, comprising:

[0031] S1. For different industries, we construct a keyword and professional terminology knowledge base based on environmental assessment industry standard documents, business needs and industry expert knowledge;

[0032] S2. Collect samples of environmental impact assessment documents from various industries, locate the corresponding chapters of the environmental impact assessment document samples according to professional terms through rules, extract candidate keywords for the chapters, calculate the similarity with professional terms, obtain approximate keywords with similarity higher than a certain threshold, and construct a keyword non-professional terminology knowledge base. For each non-professional terminology keyword, use regular expressions to establish fuzzy extraction rules. For a large number of environmental impact assessment document samples, locate the chapters of the target environmental impact assessment document and the content corresponding to the chapters according to the keyword professional terms, extract the corresponding keywords according to the fuzzy extraction rules, and if the extraction result is empty, extract the keywords as candidate keywords and calculate the similarity with the professional terms, otherwise complete the positioning. If the similarity is higher than a certain threshold, add the candidate keywords to the non-professional terminology knowledge base to improve the keyword extraction rules. The specific process of locating the candidate keywords for the chapters can be established by manually marking a knowledge base based on the chapter name, and the knowledge base covers the commonly used keywords in the environmental impact assessment documents. The specific method of improving the keyword extraction rules can be improved by manually adding the rules.

[0033] The regular expression can be: r'(monitoring|acceptance)(project|factor|indicator|content|parameter|pollution source|pollutant|control object)'. The extracted contents include: 1. Monitoring projects, monitoring factors, monitoring indicators, monitoring contents, monitoring parameters, monitoring pollution sources, monitoring pollutants, and monitoring control objects. Change all monitoring in the upper line 1 to acceptance, and all corresponding words; remove all monitoring in the upper line 1, and all corresponding words. This regular expression is a fuzzy extraction rule, and all the words that can be extracted are non-professional terminology keywords. Most words that are close or similar to professional terminology keywords will have a regular expression constructed based on the experience of a large number of documents to represent (or include) most non-professional terminology keywords.

[0034] S3. Based on the above keyword professional terminology knowledge base, keyword non-professional terminology knowledge base and keyword extraction rules, establish a (industry, professional terminology, rule, keyword) four-element relationship to generate a keyword-rule vocabulary with a tree structure;

[0035] S4. Obtain the target environmental assessment document, select the corresponding keyword professional terms from the keyword-rule thesaurus according to the industry name, locate the chapters of the target environmental assessment document and the tables and text content corresponding to the chapters, extract the corresponding keywords according to the rules in the keyword-rule thesaurus, and score the extraction results if there are any. If there are no extraction results, it is empty and is processed as if this information is missing.

[0036] The different industries mentioned are all industries in the field of environmental impact assessment, including livestock and poultry breeding industry, cement industry, electricity, heat production and supply industry, etc.

[0037] The step S1 constructs a keyword professional terminology knowledge base, traverses the environmental impact assessment industry standard documents, extracts keywords from each part of the text content and all tables in the corresponding appendix, and screens the keywords based on business needs and industry expert knowledge.

[0038] The step S2 extracts candidate keywords for the chapter, and the chapter content includes tables and texts. For the table, all words are extracted from the table in order as candidate keywords; for the text, the summary paragraph is extracted and segmented to obtain candidate keywords. The similarity between the candidate keywords and the keyword professional terms is calculated, and the candidate keywords above a certain threshold are screened out as keyword non-professional terms.

[0039] The keyword professional terms and the keyword non-professional terms are in a one-to-many relationship.

[0040] The extraction rules in step S2 include table form and text form, and different extraction rules are established for the table and text respectively.

[0041] The step S4 scores the extraction results, including:

[0042] For each keyword term, three scores are agreed upon: 1, 0.8, and 0.5, which respectively represent keywords that must be tested, keywords that need to be tested, and keywords that are icing on the cake. The score here represents the importance of the keyword.

[0043] For keywords that are not professional terms, the keyword extraction results corresponding to each extracted keyword professional term are scored. It is agreed that if the keyword and the professional term are completely consistent, it is 1 point, if the keyword and the professional term are similar (with different characters) and there is no ambiguity, it is 0 point, if the keyword and the professional term are different and there is ambiguity, it is -1 point. The score here represents the compliance of the keyword.

[0044] Calculate the keyword weight. Each keyword has two scores: the score of the keyword's professional terminology and the score of the extracted non-professional terminology keyword. Multiply these two scores to get the weight of each keyword and get the weight matrix W.

[0045] For the target environmental assessment document to be scored, check whether each extracted keyword appears in the relevant chapter. If it appears, it will be scored 1, and if it does not appear, it will be scored 0, and the score matrix S is obtained.

[0046] For each keyword, the dot product of the weight matrix W and the score matrix S is calculated to obtain the final total score. This total score is the standardization score of the target environmental impact assessment document. The more scores are greater than 0, the better the standardization, and the more scores are less than 0, the worse the standardization.

[0047] The following is a detailed description of the present invention using an environmental impact assessment document for a steel drum production and processing project as an example:

[0048] The environmental impact assessment document for the steel drum production and processing project belongs to the metal products processing and manufacturing industry (22_067). According to the evaluation level, it is necessary to extract two indicators, namely, the atmospheric evaluation level and the atmospheric pollutant proportion, from the environmental impact assessment document.

[0049] Atmospheric assessment level: The atmospheric environmental impact assessment work is divided into levels one, two and three based on conditions such as the standard emissions of major pollutants, the complexity of the surrounding terrain, and the local atmospheric environmental quality standards.

[0050] Air pollutant standard ratio: the ratio of the maximum landing concentration of pollutants in air pollutant emissions to the standard concentration.

[0051] like Figure 2 As shown, this embodiment includes the following steps:

[0052] Step S1, downloading and reading environmental impact assessment documents (reports / forms);

[0053] Step S2: Verify the validity of the data. If the data is invalid, it will be discarded and the current process will end; if the data is valid, proceed to the next step;

[0054] Step S3, extracting the chapters of the information to be extracted. If the chapters corresponding to the information are not extracted, then the full text is extracted; if the chapters corresponding to the information are extracted, then the corresponding chapters are extracted;

[0055] Step S4, using regular expressions to establish fuzzy extraction rules, establish (professional terms, rules, keywords) relationships, extract the standard rate of air pollutants and the atmospheric evaluation grade in the table, if no information is extracted, re-extract the standard rate of air pollutants and atmospheric evaluation grade information in the full text;

[0056] Step S5: Extract and output the atmospheric pollutant percentage and atmospheric evaluation grade results. The specific results are as follows: Figure 3 and Figure 4 shown.

[0057] The following is a system embodiment corresponding to the above method embodiment. This embodiment can be implemented in conjunction with the above embodiment. The relevant technical details mentioned in the above embodiment are still valid in this embodiment. In order to reduce repetition, they are not repeated here. Accordingly, the relevant technical details mentioned in this embodiment can also be applied in the above embodiment.

[0058] The present invention also provides an environmental impact assessment field semantic knowledge base application system, the system comprising:

[0059] Keyword and professional terminology knowledge base construction module, which constructs keyword and professional terminology knowledge base for different industries based on environmental assessment industry standard documents, business needs and industry expert knowledge;

[0060] The keyword non-professional terminology knowledge base construction module collects samples of environmental impact assessment documents from various industries, locates the corresponding chapters of the environmental impact assessment document samples according to the professional terminology through rules, extracts chapter keywords, calculates the similarity with the professional terminology, obtains approximate keywords with a similarity higher than a certain threshold, and constructs a keyword non-professional terminology knowledge base;

[0061] The tree-structured keyword building module uses regular expressions to establish fuzzy extraction rules for each non-professional term keyword, establishes a ternary relationship of (professional term, rule, keyword), and generates a tree-structured keyword-rule vocabulary;

[0062] The extraction rule construction module uses regular expressions to establish fuzzy extraction rules for each non-professional terminology keyword. For a large number of environmental impact assessment document samples, the chapters of the target environmental impact assessment document and the (table and text) content corresponding to the chapters are located based on the keyword professional terminology, and the corresponding keywords are extracted according to the fuzzy extraction rules. If the extraction result is empty, the chapter keywords (table and text content) are extracted and the similarity with the professional terminology is calculated. If the similarity is higher than a certain threshold, the chapter keywords are added to the non-professional terminology knowledge base to improve the keyword extraction rules.

[0063] A tree-structured keyword building module, which establishes a (professional term, rule, keyword) ternary relationship based on the above keyword professional term knowledge base, keyword non-professional term knowledge base and keyword extraction rules, and generates a tree-structured keyword-rule vocabulary;

[0064] The normative evaluation module obtains the target environmental impact assessment document, selects the corresponding keyword professional terms from the keyword-rule vocabulary according to the industry name, locates the chapters of the target environmental impact assessment document and the tables and text content corresponding to the chapters, extracts the corresponding keywords according to the rules in the keyword-rule vocabulary, and scores the extraction results if there are any, and evaluates the normativeness of the environmental impact assessment document.

Claims

1. A method for constructing a semantic knowledge base in the field of environmental impact assessment, characterized in that: include: Step 1: For each industry category, construct a professional terminology knowledge base corresponding to the industry category based on business needs and industry expert knowledge; Step 2: Collect samples of environmental impact assessment documents belonging to various industry categories, locate the target chapters of the environmental impact assessment document samples through rules, extract multiple candidate keywords in the target chapters, calculate the similarity between each candidate keyword and professional terms, and collect keywords with similarity higher than the threshold to form a non-professional terminology knowledge base. For each non-professional terminology keyword in the non-professional terminology knowledge base, establish fuzzy extraction rules through regular expressions; Step 3: Based on the professional terminology knowledge base, the non-professional terminology knowledge base and the fuzzy extraction rule, a four-element relationship among industry category, professional terminology, rule and keyword is established, and a keyword-rule vocabulary based on a tree structure is generated based on the four-element relationship; Step 4: Obtain the target environmental impact assessment document, select corresponding professional terms from the keyword-rule vocabulary according to the industry category of the target environmental impact assessment document, locate the target chapter of the target environmental impact assessment document and the tables and text content corresponding to the target chapter, extract keywords from the target chapter and the tables and text content corresponding to the target chapter according to the rules in the keyword-rule vocabulary, score the extraction results, and evaluate the standardization of the target environmental impact assessment document according to the scoring results.

2. The method for constructing a semantic knowledge base in the field of environmental impact assessment according to claim 1, characterized in that: This step 2 includes a rule updating step. For the environmental impact assessment document sample, the target chapter of the environmental impact assessment document sample is located through the professional terms in the professional terminology knowledge base, and the target chapter is extracted according to the fuzzy extraction rule. If the extraction result is empty, multiple alternative keywords in the target chapter are extracted, and the similarity between each alternative keyword and the professional term is calculated. The keywords with a collection similarity higher than a threshold are combined to form a non-professional terminology knowledge base and the fuzzy extraction rule is updated.

3. The method for constructing a semantic knowledge base in the field of environmental impact assessment according to claim 1, characterized in that: This industry category includes livestock and poultry breeding industry, cement industry, electricity, heat production and supply industry.

4. The method for constructing a semantic knowledge base in the field of environmental impact assessment according to claim 1, characterized in that: This step 1 specifically includes: traversing the environmental impact assessment industry standard documents of various industry categories, extracting keywords from each part of the text content and all tables in the corresponding appendix, screening the keywords based on business needs and industry expert knowledge, and forming a professional terminology knowledge base based on the screening results.

5. The method for constructing a semantic knowledge base in the field of environmental impact assessment according to claim 1, characterized in that: The target chapters in step 2 include tables and texts. For tables, all words are extracted in order from the tables as candidate keywords. For texts, summary paragraphs are extracted and segmented to obtain candidate keywords. The fuzzy extraction rules include table form and text form, and different extraction rules are established for the table and text respectively.

6. A semantic knowledge base construction system in the field of environmental impact assessment, characterized in that: include: Module 1: For each industry category, construct a professional terminology knowledge base corresponding to the industry category based on business needs and industry expert knowledge; Module 2: Collect samples of environmental impact assessment documents belonging to various industry categories, locate the target chapters of the environmental impact assessment document samples through rules, extract multiple candidate keywords in the target chapters, calculate the similarity between each candidate keyword and professional terms, and collect keywords with similarity higher than the threshold to form a non-professional terminology knowledge base. For each non-professional terminology keyword in the non-professional terminology knowledge base, establish fuzzy extraction rules through regular expressions; Module 3: establishing a four-element relationship among industry categories, professional terms, rules and keywords based on the professional terminology knowledge base, the non-professional terminology knowledge base and the fuzzy extraction rule, and generating a keyword-rule vocabulary based on a tree structure based on the four-element relationship; Module 4: Obtain the target environmental impact assessment document, select corresponding professional terms from the keyword-rule vocabulary according to the industry category of the target environmental impact assessment document, locate the target chapter of the target environmental impact assessment document and the tables and text content corresponding to the target chapter, extract keywords from the target chapter and the tables and text content corresponding to the target chapter according to the rules in the keyword-rule vocabulary, score the extraction results, and evaluate the standardization of the target environmental impact assessment document according to the scoring results.

7. The system for constructing a semantic knowledge base in the field of environmental impact assessment according to claim 6, characterized in that: The module 2 includes a rule updating module, which locates the target chapter of the environmental assessment document sample through the professional terms in the professional terminology knowledge base, and extracts the target chapter according to the fuzzy extraction rule. If the extraction result is empty, multiple alternative keywords in the target chapter are extracted, and the similarity between each alternative keyword and the professional term is calculated. The keywords with a collection similarity higher than a threshold are combined to form a non-professional terminology knowledge base and update the fuzzy extraction rule.

8. The system for constructing a semantic knowledge base in the field of environmental impact assessment according to claim 6, characterized in that: This industry category includes livestock and poultry breeding industry, cement industry, electricity, heat production and supply industry.

9. The system for constructing a semantic knowledge base in the field of environmental impact assessment according to claim 6, characterized in that: This module 1 specifically includes: traversing the environmental impact assessment industry standard documents of various industry categories, extracting keywords from each part of the text content and all tables in the corresponding appendix, screening the keywords based on business needs and industry expert knowledge, and forming a professional terminology knowledge base based on the screening results.

10. The system for constructing a semantic knowledge base in the field of environmental impact assessment according to claim 6, characterized in that: The target chapters in this module 2 include tables and texts; for tables, all words are extracted from the tables in order as candidate keywords; for texts, summary paragraphs are extracted and segmented to obtain candidate keywords; The fuzzy extraction rules include table form and text form, and different extraction rules are established for the table and text respectively.

Images