TWI850189B - Inquiry system for greenhouse gas emission factors - Google Patents

Abstract

An inquiry system for greenhouse gas (GHG) emission factors includes a coefficient database, a searching unit and an artificial intelligence (AI) learning unit. A data model in the coefficient database is established by standardizing at least GHG emission factor management tables of different formats according to international GHG inventory standard(s). The searching unit searches from the data model for candidate emission factor(s) as a searching result according to a searching condition and relevance weight(s) and provides the searching result for the user to select. The AI learning unit calculates and updates the relevance weight(s) according to the user’s selection by an AI algorithm. Therefore, it is possible for the user to select a more proper GHG emission factor more easily.

Description

Greenhouse Gas Emission Factor Query System

The present invention relates to a carbon inventory technology, and in particular to a greenhouse gas emission coefficient inquiry system.

When companies want to estimate greenhouse gas emissions, they usually need to first select an appropriate greenhouse gas emission factor, such as but not limited to the carbon emission factor. The selection of greenhouse gas emission factors needs to take into account many factors, such as but not limited to product or raw material category, geographical location and year. Therefore, even professionals with a certain degree of professional knowledge and experience can easily choose an inappropriate greenhouse gas emission factor, resulting in an incorrect estimate of greenhouse gas emissions.

In addition, the greenhouse gas emission coefficient is generally selected from the greenhouse gas emission calculation criteria (such as but not limited to the greenhouse gas emission coefficient management table of the Republic of China) formulated by the country or region where the enterprise is located in accordance with the international greenhouse gas inventory standards (such as but not limited to ISO14064-1:2018). If the country or region where the enterprise is located has not formulated calculation criteria, it is necessary to select from the calculation criteria formulated by neighboring countries or regions. The formats of greenhouse gas emission coefficient management tables formulated by different countries or regions are not the same, which invisibly increases the difficulty of selecting greenhouse gas emission coefficients.

In order to overcome the problems of the prior art, the purpose of the present invention is to provide a greenhouse gas emission coefficient inquiry system so that users can more conveniently and easily select a more suitable greenhouse gas emission coefficient.

The present invention provides a greenhouse gas emission coefficient query system according to an embodiment, comprising: a coefficient database for storing a data model, the data model is based on the international greenhouse gas inventory standard, and is at least a collection of data from a plurality of greenhouse gas emission coefficient tables in different formats, and includes a plurality of candidate emission coefficients, candidate conditions corresponding to each candidate emission coefficient, and at least one correlation weight value corresponding to each candidate emission coefficient, each candidate condition The data model comprises a plurality of condition parameters; a search unit is connected to the coefficient database and is used to: obtain a query condition, the query condition comprises a plurality of different query tags; according to the query condition, search for at least one candidate emission coefficient that meets the query condition from the data model and obtain the correlation weight values corresponding to each candidate emission coefficient that meets the query condition, wherein the candidate condition corresponding to each candidate emission coefficient that meets the query condition is at least partially the same as the query condition; according to each obtained The correlation weight value is used to filter out at least one candidate emission coefficient from the candidate emission coefficients that meet the requirements as a query result; and the query result is output for selection by a user; an artificial intelligence learning unit is connected to the coefficient database and is used to: obtain a coefficient selection result, the coefficient selection result points to the query condition and a target emission coefficient, and the target emission coefficient is the candidate emission coefficient selected by the user from the query result; and Using an artificial intelligence algorithm, a correlation weight value corresponding to the correlation between the query condition and the target emission coefficient is calculated according to the coefficient selection result to give the target emission coefficient to update the data model; and a graphical user interface is connected to the search unit and the artificial intelligence learning unit, and is used for the user to input data to form the query condition, and is used to present the query result to the user for selection to generate the coefficient selection result.

Optionally, the data model further includes different keyword tags, each of which is associated with at least one of the conditional parameters, and the query system further includes: a graphical user interface, which is in communication with the search unit and the artificial intelligence learning unit and is used to present the query results for the user to choose, thereby generating the coefficient selection results for the artificial intelligence learning unit, and for the user to input a part of a keyword; and a recommendation unit, which is in communication with the graphical user interface, the search unit and the artificial intelligence learning unit. The search unit is communicatively connected to the coefficient database and is used to search for matching keyword tags from the data model as keyword recommendations based on the part of the keyword, and present the keyword recommendations on the graphical user interface for the user to select. Then, based on the user's selection of the keyword recommendation, the data model is searched for related condition parameters to generate at least a part of the query condition for the search unit, wherein each matching keyword tag includes the part of the keyword.

Optionally, the data model further includes different keyword tags, each of which is associated with at least one of the conditional parameters, and the query system further includes: a graphical user interface, which is in communication with the search unit and the artificial intelligence learning unit and is used to present the query results for the user to choose, thereby generating the coefficient selection results for the artificial intelligence learning unit, and for the user to input a keyword; and a recommendation unit, which is in communication with the graphical user interface. The interface, the search unit and the coefficient database are communicatively connected, and are used to search for matching keyword tags from the data model based on the keyword, and find conditional parameters associated with the matching keyword tags from the data model as query tag recommendations, and then present the query tag recommendations on the graphical user interface for the user to select, and then generate at least a portion of the query conditions for the search unit based on the user's selection of the query tag recommendation.

Optionally, the data model further includes different keyword tags, each of which is associated with at least one of the conditional parameters, and the query system further includes: a graphical user interface, which is communicated with the search unit and the artificial intelligence learning unit and is used to present the query results for the user to select, thereby generating the coefficient selection results for the artificial intelligence learning unit, and for the user to input a text string; and a recommendation unit, which is communicated with the graphical user interface, the search unit and the coefficient database and is used to search for keyword tags appearing in the text string from the data model, and then find the conditional parameters associated with each of the searched keyword tags from the data model to generate at least a part of the query conditions for the search unit.

Optionally, the data model further includes different keyword tags, each of which is associated with at least one of the conditional parameters, and the query system further includes: a graphical user interface, which is connected to the search unit and the artificial intelligence learning unit, and is used to present the query results for the user to choose, thereby generating the coefficient selection results for the artificial intelligence learning unit, and for the user to input a file thereon; an identification A unit is communicatively connected to the graphical user interface and is used to identify or extract the text content in the file; and a recommendation unit is communicatively connected to the identification unit, the search unit and the coefficient database and is used to search for keyword tags appearing in the text content from the data model, and then find out the condition parameters associated with each of the searched keyword tags from the data model to generate at least a part of the query condition for the search unit.

Optionally, the recommendation unit first translates the data input on the graphical user interface using natural language processing (NLP) technology and large language model (LLM) technology, and then searches for keyword tags.

Optionally, each of the candidate conditions includes a main condition part and an additional condition part, and the recommendation unit selects each of the condition parameters from each of the main condition parts.

Optionally, the recommendation unit searches using natural language processing technology and generates at least a portion of the query condition using large language model technology.

Optionally, the data model is established using natural language processing technology and large language model technology based on data from the international greenhouse gas inventory standards and at least the greenhouse gas emission factor tables.

Optionally, the greenhouse gas emission coefficient tables are developed by different countries, regions or companies.

Optionally, the artificial intelligence learning unit estimates the relevance weight value using an algorithm based on reinforcement learning (RL).

Optionally, the search unit, the artificial intelligence learning unit and the recommendation unit are implemented by at least one processor. Alternatively, the search unit, the artificial intelligence learning unit, the recommendation unit and the identification unit are implemented by at least one processor.

Optionally, the coefficient database is implemented in a memory.

Alternatively, the coefficient database, the artificial intelligence learning unit, the search unit and the recommendation unit are arranged on a server-side server, and the graphical user interface is provided by an application installed on the user's computer device, and the computer device can be connected to the server-side server. Alternatively, the coefficient database, the artificial intelligence learning unit, the search unit and the recommendation unit are arranged on a server-side server, and the graphical user interface is provided by an application installed on the server-side server. Alternatively, the coefficient database, the artificial intelligence learning unit, the search unit, and the recommendation unit are located on a server-side server, and the graphical user interface is provided by an application installed on the server-side server and displayed on the user's computer device via a web page, and the computer device can be connected to the web page.

As can be seen from the above, the greenhouse gas emission coefficient query system provided by the present invention integrates greenhouse gas emission coefficient tables of different formats into a data model, allowing users to query coefficients formulated by different units (such as but not limited to countries, regions or enterprises) using only a single query tool, thereby reducing the complexity of selecting coefficients; through artificial intelligence technology, the results selected by users from the search results are statistically analyzed as the basis for providing coefficient recommendations in the future, so as to improve the accuracy of coefficient recommendations; by providing an interactive guide By providing a guidance function, the user can easily obtain keyword recommendations by inputting a partial keyword, so as to enhance the convenience of setting the query conditions; by providing an interactive guidance function, the user can easily obtain parameter recommendations by inputting a keyword, so as to enhance the convenience of setting the query conditions; and by providing a text recognition function, the user can easily obtain coefficient recommendations (i.e., query results) by inputting a text string or a file (such as but not limited to the product manual). In this way, even non-professionals can use the query system of the present invention to query more suitable greenhouse gas emission coefficients.

Please refer to FIG. 1 to FIG. 7 , the query system provided by the present invention according to one embodiment is suitable for querying greenhouse gas emission coefficients (such as but not limited to carbon emission coefficients) to provide the queryer (i.e., the user) with more accurate and more convergent query results. The query system includes at least one storage (not shown), at least one processor (not shown) and a graphical user interface 30, the aforementioned storage and the aforementioned processor are electrically connected to each other, and the graphical user interface 30 is connected to the aforementioned processor in communication.

The aforementioned storage includes a coefficient database 11, and the coefficient database 11 can store a data model. The data model is established by, for example but not limited to, utilizing natural language processing technology and large language model technology, based on international greenhouse gas inventory standards, at least aggregating data from multiple greenhouse gas emission coefficient tables in different formats, and includes multiple candidate emission coefficients, at least one candidate condition corresponding to each candidate emission coefficient, and at least one correlation weight value corresponding to each candidate emission coefficient, and each candidate condition includes multiple condition parameters. A correlation weight value represents the correlation (i.e., the degree of correlation) between a set of condition parameters (i.e., a set of query labels for a query condition) and a candidate emission coefficient (i.e., a target emission coefficient). These greenhouse gas emission coefficient tables are formulated by different countries, regions, or companies.

In this embodiment or other embodiments, the data model may optionally further include different keyword tags, each keyword tag being associated with at least one conditional parameter. For example, all keyword tags may be independent of each other; or, at least two of the keyword tags are synonyms of each other, such as the synonym of natural gas is natural gas; or, at least two of the keyword tags are similar words to each other, such as the similar words of natural gas are liquefied petroleum gas and gas; or, at least one of the keyword tags is a translation of another keyword tag in a different language (for example, the English translation of "glass" is "Glass") or a font conversion (for example, the conversion between traditional Chinese characters and simplified Chinese characters); or, any combination of the aforementioned examples.

In this embodiment or other embodiments, the data model includes, for example but not limited to, at least one associated table. For example, Table 1 presents a portion of a table, including multiple conditional parameters (i.e., "coefficient source", "state", "country", "region", "emission source form", "industrial category", "industrial process", "product or raw fuel material classification", "product or raw fuel material name", "greenhouse gas type", " _CO2 coefficient" and "coefficient unit") and their parameter contents. However, the present invention is not limited to this.

Table I Coefficient source IPCC Volume 3, Chapter 2, Table 2.4 IPCC Volume 3, Chapter 2, Formula 2.13 IPCC Volume 3, Chapter 2, Table 2.6 IPCC Volume 3, Chapter 2, Table 2.6 IPCC Volume 3, Chapter 2, Table 2.6 IPCC Volume 3, Chapter 2, Table 2.6 IPCC Volume 3, Chapter 2, Table 2.6 state – Complete Complete – – – – nation – Complete Complete – – – – Region – Complete Complete – – – – Emission source form 1.3 Process emission sources 1.3 Process emission sources 1.3 Process emission sources 1.3 Process emission sources 1.3 Process emission sources 1.3 Process emission sources 1.3 Process emission sources Industry Category Mining Industry Mining Industry Mining Industry Mining Industry Mining Industry Mining Industry Mining Industry Industrial Process Lime Process Glass Processing Glass Processing Glass Processing Glass Processing Glass Processing Glass Processing Classification of products or raw materials Raw materials product product product product product product Product or raw material name Hydraulic lime Glass Float Glass Container(Flint) Container(Amber/Green) Fiberglass(E-glass) Fiberglass(Insulation) Types of greenhouse gases CO ₂ CO ₂ CO ₂ CO ₂ CO ₂ CO ₂ CO ₂ CO ₂ coefficient 0.5887500000 0.1988095238 0.7700000000 0.7700000000 0.7700000000 0.6966666667 0.9166666667 Coefficient unit Metric ton CO ₂ / Metric ton Metric ton CO ₂ / Metric ton Metric ton CO ₂ / Metric ton Metric ton CO ₂ / Metric ton Metric ton CO ₂ / Metric ton Metric ton CO ₂ / Metric ton Metric ton CO ₂ / Metric ton

The processor at least includes a search unit 21 and an artificial intelligence learning unit 22.

The search unit 21 can be connected to the graphical user interface 30 and can obtain a query condition, and then search for candidate emission coefficients that meet the query condition from the data model, obtain the corresponding relevance weight values of each candidate emission coefficient that meets the query condition from the data model, and filter out the candidate emission coefficients from the candidate emission coefficients that meet the query condition according to the obtained relevance weight values, and finally output the query result for the user to select. The candidate conditions corresponding to each candidate emission coefficient that meets the query condition include the query condition.

The artificial intelligence learning unit 22 can communicate with the coefficient database 11 and the graphical user interface 30, and can obtain a coefficient selection result from the graphical user interface 30, and use an artificial intelligence algorithm (such as but not limited to an algorithm based on enhanced learning) to calculate the correlation weight value corresponding to the correlation between the query condition and the target emission coefficient according to the coefficient selection result to update the data model. The coefficient selection result points to the query condition and the target emission coefficient, and the target emission coefficient is the candidate emission coefficient selected by the user from the query result.

The graphical user interface 30 allows the user to input data for forming a query condition, and can present the query result for the user to select to generate a coefficient selection result.

The following is an exemplary description of the query method by which the query system provides query results.

First, as shown in step S11, the search unit 21 obtains a query condition. For example, the query condition includes multiple different query tags, such as "1.3 Process Emission Source", "Mining Engineering", "Glass Process", "Taiwan" and "2021" as shown in FIG5H.

Then, as shown in step S13, the search unit 21 searches for at least one candidate emission coefficient that meets the query condition from the data model. In this embodiment or other embodiments, at least two of the condition parameters corresponding to the candidate emission coefficient that meets the query condition are the same as at least two of the query tags of the query condition.

Next, as shown in step S15, the search unit 21 obtains each correlation weight value corresponding to each greenhouse gas emission coefficient from the data model.

Then, as shown in step S17, the search unit 21 selects the candidate emission coefficient as the query result from the candidate emission coefficients that meet the requirements according to the obtained relevance weight value. The screening condition is, for example but not limited to, selecting the candidate emission coefficients corresponding to the first N larger relevance weight values according to the order of the relevance weight values, or selecting the candidate emission coefficient corresponding to the relevance weight value greater than or equal to a preset threshold value. N is a positive integer greater than 0 and is a preset value.

Finally, as shown in step S19, the search unit 21 outputs the query result to the GUI 30 to be presented on the GUI 30 for the user to select a candidate emission factor as the target emission factor.

Since the data model covers data from greenhouse gas emission coefficient tables in different formats, users do not have to look up coefficients from these greenhouse gas emission coefficient tables in different formats one by one, thus greatly saving time and manpower.

Furthermore, when the user selects the target emission coefficient, the query system will collect statistics on the coefficient selection results to optimize the query results, as shown below. The coefficient selection results refer to the results selected by the user from the query results.

First, as shown in step S21, the artificial intelligence learning unit 22 receives the coefficient selection result from the graphical user interface 30. The coefficient selection result points to the query condition and the target emission coefficient.

Next, as shown in step S23, the artificial intelligence learning unit 22 calculates a new relevance weight value for the target emission coefficient based on the coefficient selection result using an artificial intelligence algorithm (such as but not limited to an algorithm based on enhanced learning). The relevance weight value represents the relevance between the query condition and the target emission coefficient.

Finally, as shown in step S25, the new relevance weight value is updated to the data model to replace the old relevance weight value corresponding to the target emission factor.

By optimizing the query results described above, the query results obtained by subsequent users will be more accurate than the query results obtained by previous users.

In the present invention, the processor of the query system may optionally further include a recommendation unit 23 or may optionally further include a recommendation unit 23 and an identification unit 24. The recommendation unit 23 may be connected to the graphical user interface 30, the search unit 21 and the coefficient database 11 to provide keyword recommendations to the user through interactive guidance, thereby gradually completing the setting of the query conditions. The recommendation unit 23 may, for example but not limited to, search in the coefficient database 11 using natural language processing technology, and generate at least a part of the query conditions using large language model technology, and then provide them to the search unit 21. The recognition unit 24 can be in communication with the GUI 30 and the recommendation unit 23 to recognize and capture the text of the data input on the GUI 30 and then provide it to the recommendation unit 23.

In the present invention, the query conditions can be generated in step S11 in different ways. The following examples are given to illustrate. However, the present invention is not limited to these examples.

＜Example 1＞

The graphical user interface 30 can provide multiple fields (such as but not limited to drop-down menus) at one time or step by step to the user to set different query tags (such as but not limited to "1.3 Process Emission Sources", "Mining Engineering", "Glass Process", "Taiwan" and "2021", etc.) to form query conditions for the search unit 21.

＜Example 2＞

The graphical user interface 30 may provide at least one field for the user to input a portion of a keyword, and through interactive guidance, the user may obtain keyword recommendations to gradually complete the setting of the query condition. The following will refer to FIG. 4 to exemplarily illustrate the method of using a portion of a keyword to generate at least a portion of the query condition.

First, in step S31, when the user enters a part (e.g., “玻璃”) of a keyword (e.g., but not limited to, “玻璃”) in a field provided by the graphical user interface 30 (as shown in FIG. 6A ), the recommendation unit 23 can receive the part of the keyword from the graphical user interface 30.

Next, in step S32, the recommendation unit 23 can search for a matching keyword tag (such as but not limited to "glass") from the data model as a keyword recommendation based on the keyword part by comparing it with the data of the data model, and output the keyword recommendation in step S33 to be presented on the graphical user interface 30 (as shown in FIG. 5A ) for the user to select. The matching keyword tag includes the keyword part.

Then, the recommendation unit 23 may receive the user's selection of keyword recommendation (e.g., "glass") from the graphical user interface 30 (as shown in FIG. 5B ) in step S34, and based on the selection of keyword recommendation, search for related condition parameters (e.g., but not limited to, "glass", "float glass", "container (Flint)", "container (amber)", and "fiber glass (E glass)") from the data model as query label recommendations, and then output this query label recommendation to the graphical user interface 30 (as shown in FIG. 5B ) for the user to select in step S35.

Subsequently, the recommendation unit 23 may receive the user's selection of query label recommendations (e.g., "float glass") from the graphical user interface 30 in step S36, and based on the selection of query label recommendations, search for a set of related conditional parameters (e.g., but not limited to, "1.3 process emission sources", "mining engineering", and "glass process") from the data model as query label combination recommendations, and output the query label combination recommendations to the graphical user interface 30 for the user to select, as shown in FIG5C.

Finally, the recommendation unit 23 sets the user's selection of the query tag combination recommendation as at least part of the query condition in step S37, as shown in FIG. 5D .

In this example or other examples, the graphical user interface 30 may further provide fields for the user to set other query tags, such as "Taiwan" and "2021", to make the query conditions more complete, as shown in Figures 5E to 5H.

In this example, if the user directly inputs a keyword in the GUI 30, steps S31 to S33 may be omitted.

In this example, before executing step S34, the recommendation unit 23 may first use natural language processing technology and large language model technology with the data in the data model to translate keywords, search for synonyms or similar words related to the keywords, convert fonts, or perform any combination of the foregoing.

＜Example 3＞

The graphical user interface 30 may provide a field for the user to input a text string. The text string may be, for example but not limited to, a plurality of query tags and symbols used to distinguish the query tags, or a string of text extracted from an article, or a string of other text containing at least one query tag. The following will refer to FIG. 6 to exemplarily illustrate how to generate at least a part of the query condition using the text string.

First, in step S41 , when the user inputs a text string on the GUI 30 , the recommendation unit 23 can obtain the text string from the GUI 30 .

Next, in step S42, the recommendation unit 23 may compare the text string with the keyword tags in the data model to search for the keyword tags appearing in the text string.

Then, the recommendation unit 23 finds the conditional parameters associated with each searched keyword tag from the data model in step S43.

Finally, the recommendation unit 23 sets the searched condition parameters as at least a part of the query condition in step S44 and provides it to the search unit 21.

In this example, before executing step S42, the recommendation unit 23 may first use natural language processing technology and large language model technology with the data in the data model to translate keywords, search for synonyms or similar words related to the keywords, convert fonts, or perform any combination of the foregoing.

＜Example 4＞

The graphical user interface 30 may provide a field for the user to input a file, such as but not limited to the user manual of the product. The file format may be, for example but not limited to, PDF, Word file format, Excel file format, or image file (such as but not limited to JPG, PNG, or TIFF). The following will refer to FIG. 7 to exemplarily illustrate how to generate at least a portion of the query condition using a file.

First, in step S51, when the user inputs a file on the GUI 30, the identification unit 24 may receive the file from the GUI 30.

Next, in step S52, the recognition unit 24 may recognize or extract the text content in the file to provide it to the recommendation unit 23. For example, the text content of the PDF file may be recognized using optical character recognition (OCR) technology.

Then, in step S53, the recommendation unit 23 may compare the text content of the file with the keyword tags in the data model to search for the keyword tags appearing in the text content.

Then, in step S54, the recommendation unit 23 finds conditional parameters associated with each searched keyword tag from the data model.

Finally, in step S55, the recommendation unit 23 sets the searched condition parameters as at least a part of the query condition and provides it to the search unit 21.

In this example, before executing step S53, the recommendation unit 23 may first use natural language processing technology and large language model technology with the data in the data model to translate keywords, search for synonyms or similar words related to the keywords, convert fonts, or perform any combination of the foregoing.

In this embodiment or other embodiments, the candidate condition includes a main condition part and an additional condition part, and the recommendation unit 23 selects each condition parameter from each main condition part. Taking the example of Table 1 as an example, the main condition part includes, but is not limited to, data corresponding to items such as "emission source form", "industrial category", "industrial process", "product or raw fuel material classification", "product or raw fuel material name" and "type of greenhouse gas generated", and the additional condition part includes, but is not limited to, data corresponding to items such as "state", "country" and "region". The additional condition part may even include data corresponding to items such as "year".

In the present invention, the processors and memories are arranged on a server-side server (not shown), and the graphical user interface 30 is provided by an application installed on the user's computer device (not shown), and this computer device can be connected to the server-side server; or, the processors and memories are arranged on a server-side server, and the graphical user interface 30 is provided by an application installed on this server-side server; or, the processors and memories are arranged on a server-side server, and the graphical user interface 30 is provided by an application installed on the server-side server and displayed on the user's computer device via a web page, and this computer device can be connected to this web page.

In summary, the greenhouse gas emission coefficient query system provided by the present invention integrates greenhouse gas emission coefficient tables of different formats into a data model, allowing users to query coefficients formulated by different units (such as but not limited to countries, regions or enterprises) using only a single query tool, thereby reducing the complexity of selecting coefficients; through artificial intelligence technology, the results selected by users from the search results are statistically analyzed as the basis for providing coefficient recommendations in the future, so as to improve the accuracy of coefficient recommendations; by providing an interactive guide By providing a guidance function, the user can easily obtain keyword recommendations by inputting a partial keyword, so as to enhance the convenience of setting the query conditions; by providing an interactive guidance function, the user can easily obtain parameter recommendations by inputting a keyword, so as to enhance the convenience of setting the query conditions; and by providing a text recognition function, the user can easily obtain coefficient recommendations (i.e., query results) by inputting a text string or a file (such as but not limited to the product manual). In this way, even non-professionals can use the query system of the present invention to query more suitable greenhouse gas emission coefficients.

Although the present invention is disclosed as above with the aforementioned embodiments, these embodiments are not intended to limit the present invention. Within the spirit and scope of the present invention, the changes, modifications and combinations of various embodiments are all within the scope of patent protection of the present invention. For the scope of protection defined by the present invention, please refer to the attached patent application scope.

11: Coefficient database

21:Search unit

22: Artificial Intelligence Learning Unit

23: Recommended Unit

24: Identification unit

30: Graphical User Interface

After studying the detailed description in conjunction with the following figures, other aspects of the present invention and its advantages will be discovered: Figure 1 is a functional block diagram of a greenhouse gas emission coefficient query system according to an embodiment of the present invention; Figure 2 is a flow chart of a greenhouse gas emission coefficient query method according to an embodiment of the present invention; Figure 3 is a flow chart of a method for optimizing query results according to an embodiment of the present invention; Figure 4 is a flow chart of a method for generating at least a portion of a query condition using a portion of a keyword according to an embodiment of the present invention; Figure 5A is a schematic diagram of a graphical user interface of the query system of Figure 1 to present the state of a user entering a portion of a keyword in a field; FIG. 5B is a schematic diagram of a graphical user interface of the query system of FIG. 1, showing a state of providing query tag recommendations based on a selection of keyword recommendations; FIG. 5C is a schematic diagram of a graphical user interface of the query system of FIG. 1, showing a state of providing query tag combination recommendations based on a selection of query tag recommendations; FIG. 5D is a schematic diagram of a graphical user interface of the query system of FIG. 1, showing a state of providing a field for setting a condition parameter belonging to an additional condition part; FIG. 5E is a schematic diagram of a graphical user interface of the query system of FIG. 1, showing a state of setting a condition parameter belonging to an additional condition part; FIG. 5F is a schematic diagram of a graphical user interface of the query system of FIG. 1, showing the state of providing fields for setting condition parameters belonging to the additional condition part; FIG. 5G is a schematic diagram of a graphical user interface of the query system of FIG. 1, showing the state of setting condition parameters belonging to the additional condition part; FIG. 5H is a schematic diagram of a graphical user interface of the query system of FIG. 1, showing the state of completing the setting of the query condition; FIG. 6 is a flow chart of a method for generating at least a part of the query condition using a text string according to an embodiment of the present invention; and FIG. 7 is a flow chart of a method for generating at least a part of the query condition using a file according to an embodiment of the present invention.

11: Coefficient database

21:Search unit

22: Artificial Intelligence Learning Unit

23: Recommended Units

24: Identification unit

30: Graphical User Interface

Claims (10)

A greenhouse gas emission coefficient query system includes: a coefficient database for storing a data model, the data model is based on international greenhouse gas inventory standards, and is at least a collection of data from multiple greenhouse gas emission coefficient tables in different formats, and includes different keyword tags, multiple candidate emission coefficients, candidate conditions corresponding to each candidate emission coefficient, and at least one relevance weight value corresponding to each candidate emission coefficient, each candidate condition includes multiple condition parameters, and each keyword tag is associated with at least one condition parameter; a search unit, connected to the coefficient database The communication connection is used to: obtain a query condition, the query condition including a plurality of different query tags; according to the query condition, search for at least one candidate emission factor that meets the query condition from the data model and obtain the relevance weight values corresponding to each of the candidate emission factors that meet the query condition, at least two of the condition parameters of the candidate condition corresponding to each of the candidate emission factors that meet the query condition are the same as at least two of the query tags of the query condition; according to the obtained relevance weight values, filter out at least one candidate emission factor from each of the candidate emission factors that meet the query condition as a query result; and outputting the query result for user selection; an artificial intelligence learning unit, communicating with the coefficient database and used to: obtain a coefficient selection result, the coefficient selection result pointing to the query condition and a target emission coefficient, the target emission coefficient being the candidate emission coefficient selected by the user from the query result; and using an artificial intelligence algorithm, calculating a correlation weight value corresponding to the correlation between the query condition and the target emission coefficient according to the coefficient selection result for the target emission coefficient, so as to update the data model; a graphical user interface, communicating with the The search unit is in communication with the artificial intelligence learning unit and is used for the user to input data to form the query condition, and is used to present the query result to the user for selection to generate the coefficient selection result; and a recommendation unit is in communication with the graphical user interface, the search unit and the coefficient database, and is used to search the keyword tags appearing in the input data from the data model, and then find the condition parameters associated with each of the searched keyword tags from the data model to generate at least a part of the query condition for the search unit. According to the greenhouse gas emission coefficient query system described in claim 1, the input data is a part of a keyword, the recommendation unit searches for matching keyword tags from the data model based on the part of the keyword as keyword recommendations, and presents the keyword recommendations on the graphical user interface for the user to select, and then searches for related conditional parameters from the data model based on the user's selection of the keyword recommendation, and each matching keyword tag contains the part of the keyword. According to the greenhouse gas emission coefficient query system described in claim 1, the input data is a keyword, the recommendation unit searches for a matching keyword tag from the data model based on the keyword, and finds a condition parameter associated with the matching keyword tag from the data model as a query tag recommendation, and then presents the query tag recommendation to the graphical user interface for the user to select, and then generates at least a part of the query condition to the search unit based on the user's selection of the query tag recommendation. According to the greenhouse gas emission coefficient query system described in claim 1, the input data is a text string, the recommendation unit searches for keyword tags appearing in the text string from the data model, and then finds condition parameters associated with each of the searched keyword tags from the data model, thereby generating at least a part of the query condition for the search unit. According to the greenhouse gas emission coefficient query system described in claim 1, the input data is a file, and the query system further comprises: an identification unit, which is connected to the graphical user interface and the recommendation unit, and is used to identify or extract the text content in the file to provide the recommendation unit with a search for keyword tags appearing in the text content from the data model and to find condition parameters associated with each of the searched keyword tags from the data model, thereby generating at least a part of the query condition for the search unit. A greenhouse gas emission coefficient query system according to any one of claim items 2 to 5, wherein the recommendation unit first uses natural language processing technology and large language model technology to translate the data input on the graphical user interface, and then searches for keyword tags. A greenhouse gas emission coefficient query system according to any one of claim items 2 to 5, wherein each candidate condition includes a main condition part and an additional condition part, and the recommendation unit selects each condition parameter from each main condition part. A greenhouse gas emission coefficient query system according to any one of claims 2 to 5, wherein the recommendation unit uses natural language processing technology to search and uses large language model technology to generate at least a part of the query condition. According to the greenhouse gas emission coefficient query system described in claim 1, the data model is established based on the data of the international greenhouse gas inventory standard and at least those greenhouse gas emission coefficient tables using natural language processing technology and large language model technology. According to the greenhouse gas emission coefficient query system described in claim 1, the artificial intelligence learning unit estimates the correlation weight value using an algorithm based on enhanced learning.

Images