Pesticide data repayment identification method, system, equipment and medium

Abstract

The invention relates to a pesticide data compensation and repayment identification method, system and device and a medium, and the method constructs a training set according to compensation data items in a pesticide registration process. And constructing a feature set by extracting key features in the training set. And determining that the training set data corresponding to each unique identifier meets the minimum similarity according to the similarity between the vectorized training set and the vectorized feature set. And obtaining input data similarity of the vectorized input data and the vectorized feature set, comparing the input data similarity with the minimum similarity, determining a unique identifier corresponding to the input data under the condition that the size meets the requirement, and determining the compensation cost corresponding to the experimental project based on the unique identifier. The accuracy of data compensation cost identification is improved, and the problem of low efficiency of data compensation cost identification in related technologies is solved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method, system, device and medium for identifying pesticide data compensation fees. Background Art

[0002] As a strictly regulated product, pesticides are compensated for the first data submitted for new registration, registration change, re-evaluation registration, trial use license or maintenance of existing registration in order to protect the rights and interests of registrants during the pesticide registration process. The data holder enjoys the data protection period, during which other registrants who want to cite the data must pay the data compensation fee to the data holder so that they can use these compensable data to complete the registration.

[0003] At present, in the pesticide registration process, data compensation fees are usually the largest expenditure item. The identification work is not only huge in workload, but also has high technical difficulty, resulting in low efficiency in data compensation fee identification. Therefore, the data compensation fee identification in related technologies has the problem of low efficiency. Summary of the invention

[0004] The embodiments of the present application provide a method, system, device and medium for identifying pesticide data compensation fees, so as to at least solve the problem of low efficiency in identifying data compensation fees in related technologies.

[0005] In a first aspect, an embodiment of the present application provides a method for identifying pesticide data compensation fees, wherein a training set is constructed based on the compensation data items in the pesticide registration process, wherein each compensation data item in the training set is preprocessed to determine a corresponding unique identifier; Constructing a feature set by extracting key features from the training set, wherein the key features have a high degree of similarity with the compensation data item corresponding to the unique identifier; Determining, based on the similarity between the vectorized training set and the vectorized feature set, that the training set data corresponding to each unique identifier satisfies a minimum similarity; Obtain the similarity between the vectorized input data and the input data of the vectorized feature set, compare the input data similarity with the minimum similarity, and when the size meets the requirements, determine the unique identifier corresponding to the input data, and based on the unique identifier, determine the compensation fee corresponding to the experimental project.

[0006] In one embodiment, before constructing a training set based on the compensation data items in the pesticide registration process, wherein each compensation data item in the training set is preprocessed to obtain a corresponding unique identifier, the method includes: A unique identification is established between the compensation data item and the experimental project, wherein the unique identification is used to determine the compensation fee corresponding to the compensation data item and the experimental project, and the unique identification includes a unique number and a unique identifier.

[0007] In one embodiment, each compensation data item in the training set is preprocessed to determine a corresponding unique identifier, including: By classifying each compensation data item in the training set into a corresponding unique identification category, a unique identification corresponding to each compensation data item is determined; wherein, When the number of compensation data items in the unique identification category does not meet the preset number, the compensation data items are obtained by generating synthetic data.

[0008] In one embodiment, before determining, based on the similarity between the vectorized training set and the vectorized feature set, that the training set data corresponding to each unique identifier satisfies a minimum similarity, the method includes: Preprocessing the training set, the feature set and the input data, wherein the preprocessing includes data cleaning, formatting and standardization; The preprocessed training set, the feature set and the input data are vectorized to obtain a vectorized training set, a vectorized feature set and a vectorized input data.

[0009] In one embodiment, before the similarity between the vectorized training set and the vectorized feature set, the method further comprises: The similarity between the vectorized training set and the vectorized feature set is calculated by cosine similarity, Euclidean distance or Jaccard similarity.

[0010] In one embodiment, comparing the input data similarity with the minimum similarity and determining the unique identifier corresponding to the input data when the similarity meets the requirement includes: Comparing the input data similarity with the minimum similarity; When the similarity of the input data is greater than or equal to the minimum similarity, the requirement is met, and a unique identifier corresponding to the input data is determined; When the similarity of the input data is less than the minimum similarity, the requirement is not met and the input data does not have a corresponding unique identifier.

[0011] In one embodiment, after constructing a feature set by extracting key features from the training set, the method further comprises: The training set and the feature set are updated, wherein the updating includes automatic updating and scheduled updating.

[0012] In a second aspect, an embodiment of the present application provides a pesticide data compensation fee identification system, the system comprising a training set construction module, a feature set construction module, a minimum similarity module and a compensation fee determination module, wherein: The training set construction module is used to construct a training set according to the compensation data items in the pesticide registration process, wherein each compensation data item in the training set is pre-processed to determine the corresponding unique identifier; The feature set building module is used to build a feature set by extracting key features in the training set, wherein the key features have a high similarity with the compensation data item corresponding to the unique identifier; The minimum similarity module is used to determine that the training set data corresponding to each unique identifier meets the minimum similarity according to the similarity between the vectorized training set and the vectorized feature set; The compensation fee determination module is used to obtain the similarity between the vectorized input data and the input data of the vectorized feature set, compare the input data similarity with the minimum similarity, and determine the unique identifier corresponding to the input data when the size meets the requirements, and determine the compensation fee corresponding to the experimental project based on the unique identifier.

[0013] In a third aspect, an embodiment of the present application provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, a method for identifying pesticide data compensation fees as described in the first aspect above is implemented.

[0014] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for identifying pesticide data compensation fees as described in the first aspect above.

[0015] The pesticide data compensation fee identification method, system, device and medium provided in the embodiments of the present application have at least the following technical effects.

[0016] A training set is constructed based on the compensation data items in the pesticide registration process, wherein each compensation data item in the training set is determined to have a corresponding unique identifier through preprocessing. A feature set is constructed by extracting key features in the training set, wherein the key features have a high degree of similarity with the compensation data items corresponding to the unique identifier. According to the similarity between the vectorized training set and the vectorized feature set, it is determined that the training set data corresponding to each unique identifier meets the minimum similarity. The similarity between the vectorized input data and the input data of the vectorized feature set is obtained, and the size of the input data similarity and the minimum similarity is compared. When the size meets the requirements, the unique identifier corresponding to the input data is determined, and the compensation fee corresponding to the experimental project is determined based on the unique identifier. By comparing the input data similarity with the similarity of the vectorized training set and the feature set, the unique identifier corresponding to the input data is identified, and then the compensation fee corresponding to the experimental project is determined by the unique identifier, thereby improving the accuracy of data compensation fee identification and solving the problem of low efficiency in data compensation fee identification in related technologies.

[0017] Details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 It is a flow chart of a method for identifying pesticide data compensation fees; Figure 2 is a schematic diagram of the overall process structure according to an exemplary embodiment; Figure 3 is a structural block diagram of a pesticide data compensation fee identification system according to an exemplary embodiment; Figure 4 It is a structural block diagram of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0020] Obviously, the drawings described below are only some examples or embodiments of the present application. For ordinary technicians in this field, the present application can also be applied to other similar scenarios based on these drawings without creative work. In addition, it can also be understood that although the efforts made in this development process may be complicated and lengthy, for ordinary technicians in this field related to the content disclosed in this application, some changes in design, manufacturing or production based on the technical content disclosed in this application are just conventional technical means, and should not be understood as insufficient content disclosed in this application.

[0021] Reference to "embodiments" in this application means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those of ordinary skill in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0022] Unless otherwise defined, the technical terms or scientific terms involved in this application should be understood by people with ordinary skills in the technical field to which this application belongs. The words "one", "a", "a", "the" and the like involved in this application do not indicate a quantity limitation, and may indicate the singular or plural. The terms "include", "comprise", "have" and any of their variations involved in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there may be three relationships, for example, "A and / or B" can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects before and after are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0023] In the pesticide registration process, data compensation fees are usually the largest expense item. Its evaluation is not only labor-intensive but also technically difficult. The main challenges include: 1. The data lists provided by data holders are often lengthy, containing hundreds to thousands of data items. For conventional chemical pesticides, data compensation fees usually reach tens of millions or even hundreds of millions of RMB. However, not all data require compensation. New registrants need to judge the necessity and validity of the data on their own. Accurately judging the required data compensation fees will help reduce the registration expenses of new registrants.

[0024] 2. To evaluate the necessity and validity of the data in the data list, a deep technical background is required, involving knowledge in professional fields such as plant protection, residues, toxicology, and the environment, and an understanding of the cost of generating these data. However, Chinese pesticide companies, as new registration applicants, often lack this comprehensive technical background.

[0025] 3. The necessity and validity assessment of data is usually done manually, which is time-consuming and labor-intensive, and requires the evaluators to have comprehensive technical capabilities, resulting in low overall efficiency.

[0026] Currently, there is no relevant automated software tool that can improve the efficiency of data compensation cost identification in pesticide registration.

[0027] Therefore, the present application provides a method, system, device and medium for identifying pesticide data compensation fees, so as to at least solve the problem of low efficiency in data compensation fee identification in related technologies.

[0028] In a first aspect, an embodiment of the present application provides a method for identifying pesticide data compensation fees. Figure 1 It is a flow chart of a method for identifying pesticide data compensation fees. Figure 2 is a schematic diagram of the overall process structure according to an exemplary embodiment. Figure 1 and Figure 2 As shown, a method for identifying pesticide data compensation fee includes: Step S101: construct a training set according to the compensation data items in the pesticide registration process, wherein each compensation data item in the training set is pre-processed to determine a corresponding unique identifier.

[0029] Step S102: construct a feature set by extracting key features from the training set, wherein the key features have a high degree of similarity with the compensation data items corresponding to the unique identifier.

[0030] Step S103: According to the similarity between the vectorized training set and the vectorized feature set, determine whether the training set data corresponding to each unique identifier meets the minimum similarity.

[0031] Step S104, obtain the similarity between the vectorized input data and the input data of the vectorized feature set, compare the input data similarity with the minimum similarity, and when the size meets the requirements, determine the unique identifier corresponding to the input data, and based on the unique identifier, determine the compensation fee corresponding to the experimental project.

[0032] In summary, the embodiment of the present application provides a method for identifying pesticide data compensation fees, which constructs a training set based on the compensation data items in the pesticide registration process, wherein each compensation data item in the training set is determined by preprocessing to have a corresponding unique identifier. A feature set is constructed by extracting key features in the training set, wherein the key features have a high degree of similarity with the compensation data items corresponding to the unique identifier. According to the similarity between the vectorized training set and the vectorized feature set, it is determined that the training set data corresponding to each unique identifier meets the minimum similarity. The similarity between the vectorized input data and the input data of the vectorized feature set is obtained, and the size of the input data similarity and the minimum similarity is compared. When the size meets the requirements, the unique identifier corresponding to the input data is determined, and the compensation fee corresponding to the experimental project is determined based on the unique identifier. By comparing the input data similarity with the similarity between the vectorized training set and the feature set, the unique identifier corresponding to the input data is identified, and then the compensation fee corresponding to the experimental project is determined by the unique identifier, thereby improving the accuracy of data compensation fee identification and solving the problem of low efficiency in data compensation fee identification in related technologies.

[0033] In one embodiment, in step S101, a training set is constructed according to the compensation data items in the pesticide registration process, wherein before each compensation data item in the training set is preprocessed to determine a corresponding unique identifier, the method further comprises: A unique identification is established between the compensation data item and the experimental project, wherein the unique identification is used to determine the compensation fee corresponding to the compensation data item and the experimental project, and the unique identification includes a unique number and a unique identifier.

[0034] Optionally, a unique identification can be established for each compensation data item that needs data compensation and its corresponding test item. The unique identification includes a unique number and a unique identifier. A mapping relationship is established between each compensation data item and its corresponding unique identification (including a unique number and a unique identifier) ​​to ensure that each data item can accurately correspond to a specific test item. The mapping relationship is stored in a database or data structure for subsequent query and use.

[0035] Among them, according to the OPPTS (Office of Prevention, Pesticides and Toxic Substances) series of test guidelines, the unique identifier can be divided into two segments of numbers with a ".". For example, the standard OPPTS number is 850.1075, where 850 represents the test guidelines for the ecotoxicology test series, and 1075 represents the acute toxicity test for freshwater or saltwater fish. Further, the subdivision number is added to form a three-segment number by adding another "." to distinguish different test items. For example, 850.1075.1 represents the acute toxicity test of warm water freshwater fish using the static test method, including fish such as bluegill sunfish, channel catfish, common carp, fathead minnow, zebrafish, medaka, guppy, etc.

[0036] It is worth noting that in addition to the unique number, each test item can also be assigned a unique identifier, such as a letter, symbol or combination, to increase the uniqueness and readability of the identification.

[0037] The beneficial effect of establishing unique identification for compensation data items and experimental items is as follows: by setting unique numbers, a mapping relationship between compensation data items and unique identification is established to ensure that each data item can accurately correspond to a specific test item. In addition, by referring to the OPPTS series of test guidelines through unique numbers, the standardization and uniformity of numbers can be ensured, which facilitates communication and cooperation between different institutions and personnel. The test items are further subdivided through three-segment digital numbers to ensure the uniqueness and identifiability of each test item and avoid confusion and duplication.

[0038] In one embodiment, step S101, constructing a training set based on the compensation data items in the pesticide registration process, wherein each compensation data item in the training set is pre-processed to determine the corresponding unique identifier. Specifically, it includes: By classifying each compensation data item in the training set into a corresponding unique identification category, the unique identification corresponding to each compensation data item is determined; wherein, When the number of compensation data items in the unique identification category does not meet the preset number, the compensation data items are obtained by generating synthetic data.

[0039] Optionally, each compensated data item in the training set is classified into a corresponding unique identification category according to the type and characteristics of the test project. The data items in the training set are real data in the pesticide registration process, and are classified into corresponding unique numbers according to categories by technicians with professional technical backgrounds. The compensated data items in each category have a unique identification, including a unique number and a unique identifier. For example, 850.1075.1 represents an acute toxicity test for warm water freshwater fish using a static test method. The number of data items under each category should meet the minimum quantity requirement m. When the amount of real data under a category is less than m, the training set is supplemented by generating synthetic data. Large language models (LLMs) or other machine learning techniques can be used to generate qualified synthetic data based on specific prompts. Synthetic data also simulates real data.

[0040] Step S101 can ensure that each compensated data item can be accurately classified into the corresponding unique identification category, and supplement the training set by generating synthetic data when the data amount is insufficient. By generating synthetic data, it is ensured that the amount of data in each category meets the minimum quantity requirement, thereby improving the integrity and representativeness of the training set.

[0041] In one embodiment, step S102, constructing a feature set by extracting key features from a training set, wherein the key features and the compensation data items corresponding to the unique identifier have a high degree of similarity. Specifically, it includes: The feature set is constructed by extracting the key features in the training set, and the feature set is used to extract key features from the training set data classified into the unique number of the test item. The key feature extraction is summarized and summarized by technicians with professional technical background based on experience, requiring that the key features under a unique number have a high degree of similarity with the training set data items under the number, and at the same time have a distinguishable margin with the training set data items under other unique numbers.

[0042] Specifically, the feature set is mainly based on manual judgment based on expert experience, and then verified and modified in the vector database. For example, for the acute toxicity test of warm-water fish using static, static-renewal and semi-static test methods, the corresponding test item number is 850.1075.1, and the corresponding feature set covers different test methods, the Latin names of different warm-water fish species, and characteristic descriptions of acute toxicity. In this way, tests with similar costs can be classified into one category. If other test methods (such as flow water method) or different species (such as cold water fish or saltwater fish) are used, although they are all acute toxicity tests for fish, the costs are different, so they need to be classified into other test item numbers. Based on these different information, the feature set constructed for test item 850.1075.1 contains 7 data, as follows: ##850.1075.1.1## Acute toxicity, static, static renewal, semistatic, bluegill sunfish (Lepomis macrochirus); ##850.1075.1.2## Acute toxicity, static, static renewal, semistatic, channel catfish (Ictaluruspunctatus); ##850.1075.1.3## Acute toxicity, static, static renewal, semistatic, carp (Cyprinus carpio); ##850.1075.1.4## Acute toxicity, static, static renewal, semi-static, fathead carp (Pimephalespromelas); ##850.1075.1.5## Acute toxicity, static, static renewal, semistatic, guppy (Poeciliareticulata); ##850.1075.1.6## Acute toxicity, static, static renewal, semi-static, medaka (Oryzias latipes); ##850.1075.1.7## Acute toxicity, static, static renewal, semi-static, zebrafish (Danio rerio); After the above feature set is constructed, it is necessary to further verify the constructed feature set in the vector database to manually determine whether it can match accurate real data. If the matching result is not ideal, further feature optimization is required until a satisfactory matching degree is achieved. For example, the feature set of 850.1075.1 above should accurately match real data similar to the following: Samel, A. (2004) DPX-E2Y45 technical: static, acute, 96-hour LC50 for bluegill sunfish (Lepomis macrochirus). Project Nos. DUPONT / 12333, 14513, 226. Unpublished research prepared by EI du Pont deNemours And Co., Inc., 34 pages.

[0043] Collins, M. (1993) MB46513: Acute toxicity to bluegill sunfish (Lepomis macrochirus) under static renewal conditions: Final report: Laboratory Project Number: 92 / 9 / 4425: 10566 / 0492 / 6242 / 100. Unpublished research prepared by Springborn Laboratories, Inc., 61 pages.

[0044] Turner, J. (2006) DPX-E2Y45 20SC [200 g / L (W / V); 18.5% (W / W)]: Static, acute, 96-hour limit test on bluegill sunfish (Lepomis macrochirus). Project Nos. DUPONT / 18602, 16294,226. Unpublished research prepared by EI du Pont de Nemours And Co., Inc., 33 pages.

[0045] Dorgerloh, M. (1999) JAU6476 - Acute toxicity to bluegill sunfish (Lepomis macrochirus) in a static test (96 hours). Project No.: E / 2521695 / 0, DOM / 99090, 00586. Unpublished study prepared by Bayer AG Institut fuer Ruckstands-Analytik, 48 pages.

[0046] Peither, A. (2000) Acute toxicity of MTI-446 to bluegill sunfish (Lepomis macrochirus) in a 96-hour static test: Laboratory Project No. 740946: 740957. Unpublished research prepared by RCC, Ltd., 33 pages. {OPPTS 850.1075} Kern, M.; Lam, C. (2003) JAU 6476 480 SC Acute toxicity to bluegill sunfish (Lepomis macrochirus) under static renewal conditions. Project Nos. 200599, EBJAX075, J6810301. Unpublished research prepared by Bayer Corp., 34 pages.

[0047] The above-mentioned vector database is a vectorized collection of all real data (i.e., data of all real test items such as 850.1075.1, 850.1075.2, 890.1100.1, etc.). The vector database will also be continuously updated in the subsequent application process of the pesticide data compensation fee identification system.

[0048] The purpose of the training set data item having a distinguishable margin is to find the real data that the feature set matches accurately but has the lowest similarity in the vector database. This data will be used as the worst case of data matching, that is, the lowest distinguishable margin. For example, the real data that the above 850.1075.1.1 matches accurately but has the lowest similarity in the vector database is Kern, M.; Lam, C. (2003) JAU 6476 480 SC Acute toxicity to bluegill sunfish (Lepomis macrochirus), under static update conditions. Project number: 200599, EBJAX075, J6810301. Unpublished research prepared by Bayer Corp., 34 pages. The calculated similarity of this data is 75%, which is the lowest similarity among other real data that match accurately. Therefore, if the data matched by the feature set in the actual application process is 850.1075.1.1, the similarity is lower than 75%, which is considered unsuccessful, and higher than 75%, which is considered successful.

[0049] Step S102 ensures that key features can be extracted from the data items of each test item by constructing a feature set, thereby improving the standardization and uniformity of the data. The extracted key features should be highly similar to the training set data items under the same unique number, ensuring that the features can accurately represent the data items of this type of test item. The extracted key features should be distinguishable from the training set data items under other unique numbers, ensuring that data items of different categories can be correctly identified and classified.

[0050] In one embodiment, before determining, in step S103, based on the similarity between the vectorized training set and the vectorized feature set, that the training set data corresponding to each unique identifier satisfies the minimum similarity, the method further includes: Preprocess the training set, feature set and input data, where preprocessing includes data cleaning, formatting and standardization; The preprocessed training set, feature set and input data are vectorized to obtain a vectorized training set, a vectorized feature set and a vectorized input data.

[0051] Optionally, first, the training set, feature set and input data are preprocessed. The preprocessing includes data cleaning: removing irrelevant information and noise to ensure the accuracy and consistency of the data. Formatting: formatting the data into a unified standard to facilitate subsequent processing and analysis. Standardization: normalizing the data to ensure the comparability between different data items.

[0052] Secondly, the training set, feature set and input data are vectorized using a semantic vector model to convert the data into a vector form. The embodiment of the present application does not limit the type of semantic vector model used, and both open source models and closed source models can be used, and the models can be locally built or called online.

[0053] Through preprocessing and vectorization, it is possible to ensure that the training set, feature set, and input data are fully preprocessed before vectorization, thereby improving the quality of the data and the accuracy of matching. Vectorization converts the data into vector form, which is convenient for subsequent calculation and comparison.

[0054] In one embodiment, step S103, based on the similarity between the vectorized training set and the vectorized feature set, determines that the training set data corresponding to each unique identifier meets the minimum similarity. Specifically includes: Calculate the similarity between the vectorized feature set and the vectorized training set. A variety of similarity algorithms can be used, such as cosine similarity, Euclidean distance, Jaccard similarity, etc. By calculating the similarity between vectors, determine the minimum similarity that the training set data under each unique number should meet. Vectorize the input data. Calculate the similarity between the vectorized input data and the vectorized feature set. Compare the calculated similarity with the minimum similarity, the minimum similarity threshold can be 65%-80%, and the minimum similarity in the embodiment of the present application can be 75%. If the similarity is greater than or equal to the minimum similarity, it is determined that the input data should match the unique number.

[0055] For example, two new [input data A] and [input data B] are vectorized and then similarity is calculated with the vectorized [feature set]. The identified test item is the test item with the highest similarity to the features of different test items covered in the [feature set]. For example, the test items identified by [input data A] and [input data B] are both OPPTS850.1075.1, with similarities of 80% and 60% respectively. Because the [feature set] needs to be continuously updated and cannot cover all test items, the [feature set] is a high-level summary of the key features of different test items, so it has a high recognition accuracy. However, if the features of a certain test item are not included in the [feature set], a test item with the highest similarity but incorrect recognition will be recognized, so it needs to be further excluded. For example, the similarity between [input data B] and the test item OPPTS 850.1075.1 is low, which may be a wrong recognition. Therefore, in order to further determine whether the data identified based on the [feature set] is accurate, it is necessary to introduce another [training set] consisting of real data. The data of different test items in the [training set] have been vectorized and similarity calculated with the vectorized [feature set]. For example, there are 50 test items OPPTS850.1075.1 in the [training set], and the minimum similarity between these 50 data items and the [feature set] is 75%. Therefore, if the similarity between the [input data] and the [feature set] exceeds 75%, it means that the [input data] is accurately identified as OPPTS850.1075.1, such as [input data A] here. On the contrary, if it is lower than this threshold, it means that the identification is inaccurate, such as [input data B], so further manual judgment is required. One possible situation is that the test item feature corresponding to [input data B] is not included in the [feature set]. Assuming that the test item that should be accurately identified is OPPTS 850.1075.2, in this case, the [feature set] and the corresponding [training set] should be updated to ensure the subsequent accurate identification of this data.

[0056] Step S103 quantifies the similarity between the data by calculating the similarity between the vectorized feature set and the vectorized training set, providing a quantitative indicator for subsequent matching judgment. By calculating the similarity between the vectorized input data and the vectorized feature set and comparing it with the minimum similarity, it is determined whether the input data should match the unique number, thereby ensuring the classification accuracy and reliability of the input data.

[0057] In one embodiment, step S104 obtains the similarity between the vectorized input data and the input data of the vectorized feature set, compares the input data similarity with the minimum similarity, and determines the unique identifier corresponding to the input data if the similarity meets the requirements, and determines the compensation fee corresponding to the experimental project based on the unique identifier. Specifically, it includes: Compare the input data similarity with the minimum similarity; When the similarity of the input data is greater than or equal to the minimum similarity, the requirement is met and the unique identifier corresponding to the input data is determined; When the similarity of the input data is less than the minimum similarity, it does not meet the requirements and the input data does not have a corresponding unique identifier.

[0058] Optionally, obtain the similarity between the vectorized input data and the vectorized feature set. Obtain the minimum similarity threshold that the training set data under each unique number should meet. Compare the input data similarity with the minimum similarity. When the input data similarity is greater than or equal to the minimum similarity, it meets the requirements and determines the unique identifier corresponding to the input data. When the input data similarity is less than the minimum similarity, it does not meet the requirements and the input data does not have a corresponding unique identifier. Based on the unique identifier, determine the compensation fee corresponding to the experimental project.

[0059] Step S104 quantifies the similarity between the input data and the known feature set by calculating the similarity between the vectorized feature set and the vectorized input data, providing a quantitative indicator for subsequent matching judgment. By comparing the input data similarity with the minimum similarity, it is determined whether the input data is matched successfully. When the match is successful, the unique identifier corresponding to the input data is determined, and then the experimental project and its corresponding compensation fee are determined; when the match fails, it indicates that the input data has not been recognized by the current software and needs to be further optimized.

[0060] In one embodiment, after constructing the feature set by extracting key features in the training set, the method further includes: Update the training set and feature set, including automatic update and scheduled update.

[0061] Optionally, update the training set and feature set to ensure real-time response to changes in new data, timely update the training set and feature set, and improve the real-time and adaptability of the system. Scheduled updates ensure that the training set and feature set are regularly updated at predetermined time intervals to ensure the timeliness and accuracy of the data. Continuously collect new compensation data items to ensure the volume and diversity of the training set and feature set, which helps the system to continuously learn and optimize and improve the generalization ability and accuracy of the model.

[0062] In one embodiment, the input data is preprocessed to exclude as many data items that do not require data compensation as possible based on key information (such as whether it is a publicly published literature document, etc.).

[0063] Based on key information (such as whether it is publicly published literature), data items that do not require data compensation are excluded, which reduces unnecessary processing and improves system efficiency. Irrelevant information and noise are deleted to ensure the accuracy and consistency of data, reduce interference factors, and improve the efficiency and accuracy of subsequent processing.

[0064] In one embodiment, based on the results of matching the unique identifier, multiple strategies are used to further optimize the results to improve the overall accuracy.

[0065] In summary, the embodiment of the present application provides a method for identifying pesticide data compensation fees, which constructs a training set based on the compensation data items in the pesticide registration process, wherein each compensation data item in the training set is determined by preprocessing to have a corresponding unique identifier. A feature set is constructed by extracting key features in the training set, wherein the key features have a high degree of similarity with the compensation data items corresponding to the unique identifier. According to the similarity between the vectorized training set and the vectorized feature set, it is determined that the training set data corresponding to each unique identifier meets the minimum similarity. The similarity between the vectorized input data and the input data of the vectorized feature set is obtained, and the size of the input data similarity and the minimum similarity is compared. When the size meets the requirements, the unique identifier corresponding to the input data is determined, and the compensation fee corresponding to the experimental project is determined based on the unique identifier. By comparing the input data similarity with the similarity between the vectorized training set and the feature set, the unique identifier corresponding to the input data is identified, and then the compensation fee corresponding to the experimental project is determined by the unique identifier, thereby improving the accuracy of data compensation fee identification and solving the problem of low efficiency in data compensation fee identification in related technologies.

[0066] It is worth noting that the embodiment of the present application has no restrictions on the type of semantic vector model used. Open source or closed source models can be used, and the model can be built locally or called online. The practice of the embodiment of the present application adopts the local open source model BAAI / llm-embedder, which runs in CPU mode, with the hardware of Intel(R) Xeon(R) Silver 4310 CPU @2.10GHz and 24 GB of memory. When the number of data items is more than 600, the running time is about 1 minute, and the recognition accuracy is above 90%.

[0067] The intelligent identification of pesticide data compensation fees can significantly improve the efficiency of pesticide registration data compensation identification, and can continuously learn and identify more test items during actual application.

[0068] In a second aspect, an embodiment of the present application provides a pesticide data compensation fee identification system. Figure 3 FIG. 1 is a block diagram of a pesticide data compensation fee identification system according to an exemplary embodiment. Figure 3As shown, the system includes a training set construction module 310, a feature set construction module 320, a minimum similarity module 330 and a compensation fee determination module 340, wherein: A training set building module 310 is used to build a training set according to the compensation data items in the pesticide registration process, wherein each compensation data item in the training set is determined to have a corresponding unique identifier through preprocessing; A feature set building module 320 is used to build a feature set by extracting key features in a training set, wherein the key features have a high similarity with the compensation data item corresponding to the unique identifier; A minimum similarity module 330 is used to determine whether the training set data corresponding to each unique identifier satisfies the minimum similarity based on the similarity between the vectorized training set and the vectorized feature set; The compensation fee determination module 340 is used to obtain the similarity between the vectorized input data and the input data of the vectorized feature set, compare the input data similarity and the minimum similarity, and determine the unique identifier corresponding to the input data when the size meets the requirements, and determine the compensation fee corresponding to the experimental project based on the unique identifier.

[0069] In summary, the present application provides a pesticide data compensation fee identification system. By constructing a training set module 310, constructing a feature set module 320, a minimum similarity module 330 and determining a compensation fee module 340. The problem of information redundancy in the event extraction process can be significantly reduced, the accuracy of event extraction can be improved, and the problem of low accuracy of event extraction methods in related technologies can be solved. Specifically, a training set is constructed according to the compensation data items in the pesticide registration process, wherein each compensation data item in the training set is determined by preprocessing to determine the corresponding unique identifier. A feature set is constructed by extracting key features in the training set, wherein the key features have a high degree of similarity with the compensation data items corresponding to the unique identifier. According to the similarity between the vectorized training set and the vectorized feature set, it is determined that the training set data corresponding to each unique identifier meets the minimum similarity. The similarity between the vectorized input data and the input data of the vectorized feature set is obtained, and the size of the input data similarity and the minimum similarity is compared. When the size meets the requirements, the unique identifier corresponding to the input data is determined, and based on the unique identifier, the compensation fee corresponding to the experimental project is determined. By comparing the similarity of the input data with the similarity of the vectorized training set and feature set, the unique identifier corresponding to the input data is identified, and then the compensation cost corresponding to the experimental project is determined through the unique identifier, which improves the accuracy of data compensation cost identification and solves the problem of low efficiency in data compensation cost identification in related technologies.

[0070] It should be noted that the pesticide data compensation fee identification system provided in this embodiment is used to implement the above-mentioned implementation mode, and the descriptions that have been made will not be repeated. As used above, the terms "module", "unit", "subunit", etc. can implement a combination of software and / or hardware for a predetermined function. Although the device described in the above embodiment is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0071] In a third aspect, an embodiment of the present application provides an electronic device, Figure 4 FIG. 1 is a block diagram of an electronic device according to an exemplary embodiment. Figure 4 As shown, the electronic device may include a processor 41 and a memory 42 storing computer program instructions.

[0072] Specifically, the processor 41 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0073] Among them, the memory 42 may include a large-capacity memory for data or instructions. By way of example and not limitation, the memory 42 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 42 may include a removable or non-removable (or fixed) medium. Where appropriate, the memory 42 may be inside or outside the data processing device. In a specific embodiment, the memory 42 is a non-volatile memory. In a specific embodiment, the memory 42 includes a read-only memory (ROM) and a random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically alterable ROM (EAROM) or a flash memory (FLASH), or a combination of two or more of these. Under appropriate circumstances, the RAM can be a static random access memory (SRAM) or a dynamic random access memory (DRAM), wherein the DRAM can be a fast page mode dynamic random access memory (FPMDRAM), an extended data output dynamic random access memory (EDODRAM), a synchronous dynamic random access memory (SDRAM), etc.

[0074] The memory 42 may be used to store or cache various data files required for processing and / or communication, as well as possible computer program instructions executed by the processor 41 .

[0075] The processor 41 implements any one of the pesticide data compensation fee identification methods in the above embodiments by reading and executing computer program instructions stored in the memory 42 .

[0076] In one embodiment, a pesticide data compensation fee identification device may further include a communication interface 43 and a bus 40. Figure 4 As shown, the processor 41, the memory 42, and the communication interface 43 are connected via a bus 40 and communicate with each other.

[0077] The communication interface 43 is used to implement communication between the modules, devices, units and / or equipment in the embodiment of the present application. The communication port 43 can also implement data communication with other components such as: external devices, image / data acquisition equipment, databases, external storage, and image / data processing workstations.

[0078] The bus 40 includes hardware, software or both, and couples the components of a pesticide data compensation fee identification device to each other. The bus 40 includes but is not limited to at least one of the following: a data bus, an address bus, a control bus, an expansion bus, and a local bus. By way of example and not limitation, bus 40 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses or a combination of two or more of these. Where appropriate, bus 40 may include one or more buses. Although embodiments of the present application describe and illustrate a particular bus, the present application contemplates any suitable bus or interconnect.

[0079] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements a pesticide data compensation fee identification method provided in the first aspect.

[0080] The readable storage medium may include but is not limited to: a portable disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical storage device, a magnetic storage device or any suitable combination of the above.

[0081] In a possible implementation, the present invention can also be implemented in the form of a program product, which includes program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps of a pesticide data compensation fee identification method provided in the first aspect.

[0082] Among them, the program code for executing the present invention can be written in any combination of one or more programming languages, and the program code can be executed completely on the user device, partially on the user device, as an independent software package, partially on the user device and partially on a remote device, or completely on the remote device.

[0083] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0084] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A method for identifying pesticide data compensation fees, characterized in that: The method comprises: Constructing a training set according to the compensation data items in the pesticide registration process, wherein each compensation data item in the training set is preprocessed to determine a corresponding unique identifier; Constructing a feature set by extracting key features from the training set, wherein the key features have a high degree of similarity with the compensation data item corresponding to the unique identifier; Determining, based on the similarity between the vectorized training set and the vectorized feature set, that the training set data corresponding to each unique identifier satisfies a minimum similarity; Obtain the similarity between the vectorized input data and the input data of the vectorized feature set, compare the input data similarity with the minimum similarity, and when the size meets the requirements, determine the unique identifier corresponding to the input data, and based on the unique identifier, determine the compensation fee corresponding to the experimental project.

2. The method according to claim 1, characterized in that: The comparing the input data similarity with the minimum similarity, and determining the unique identifier corresponding to the input data when the similarity meets the requirement, includes: Comparing the input data similarity with the minimum similarity; When the similarity of the input data is greater than or equal to the minimum similarity, the requirement is met, and a unique identifier corresponding to the input data is determined; When the similarity of the input data is less than the minimum similarity, the requirement is not met and the input data does not have a corresponding unique identifier.

3. The method according to claim 1, characterized in that Before constructing a training set based on the compensation data items in the pesticide registration process, wherein each compensation data item in the training set obtains a corresponding unique identifier through preprocessing, the method further includes: A unique identification is established between the compensation data item and the experimental project, wherein the unique identification is used to determine the compensation fee corresponding to the compensation data item and the experimental project, and the unique identification includes a unique number and a unique identifier.

4. The method according to claim 1, characterized in that: Each compensation data item in the training set is preprocessed to determine a corresponding unique identifier, including: By classifying each compensation data item in the training set into a corresponding unique identification category, a unique identification corresponding to each compensation data item is determined; wherein, When the number of compensation data items in the unique identification category does not meet the preset number, the compensation data items are obtained by generating synthetic data.

5. The method according to claim 1, characterized in that: Before determining, based on the similarity between the vectorized training set and the vectorized feature set, that the training set data corresponding to each unique identifier satisfies the minimum similarity, the method further includes: Preprocessing the training set, the feature set and the input data, wherein the preprocessing includes data cleaning, formatting and standardization; The preprocessed training set, the feature set and the input data are vectorized to obtain a vectorized training set, a vectorized feature set and a vectorized input data.

6. The method according to claim 1, characterized in that: Before the similarity between the vectorized training set and the vectorized feature set, the method further includes: The similarity between the vectorized training set and the vectorized feature set is calculated by cosine similarity, Euclidean distance or Jaccard similarity.

7. The method according to claim 1, characterized in that: After constructing a feature set by extracting key features from the training set, the method further includes: The training set and the feature set are updated, wherein the updating includes automatic updating and scheduled updating.

8. A pesticide data compensation fee identification system, characterized in that: The system includes a training set construction module, a feature set construction module, a minimum similarity module and a compensation fee determination module, wherein: The training set construction module is used to construct a training set according to the compensation data items in the pesticide registration process, wherein each compensation data item in the training set is pre-processed to determine the corresponding unique identifier; The feature set building module is used to build a feature set by extracting key features in the training set, wherein the key features have a high similarity with the compensation data item corresponding to the unique identifier; The minimum similarity module is used to determine that the training set data corresponding to each unique identifier meets the minimum similarity according to the similarity between the vectorized training set and the vectorized feature set; The compensation fee determination module is used to obtain the similarity between the vectorized input data and the input data of the vectorized feature set, compare the input data similarity with the minimum similarity, and determine the unique identifier corresponding to the input data when the size meets the requirements, and determine the compensation fee corresponding to the experimental project based on the unique identifier.

9. An electronic device, characterized in that: The method comprises a memory and a processor, a computer program stored in the memory and executable on the processor, and the processor implements a pesticide data compensation fee identification method as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, a pesticide data compensation fee identification method as described in any one of claims 1 to 7 is implemented.

Images