A data product recommendation method, device, equipment, storage medium and product

Abstract

This application discloses a data product recommendation method, apparatus, device, storage medium, and product. The method includes: acquiring scenario demand information input by a data buyer; extracting keywords from the scenario demand information to form a target keyword set; vectorizing the target keyword set to form a first set of vectors; vectorizing text information from multiple data providers to obtain multiple second sets of vectors; calculating the similarity between the first set of vectors and each of the second set of vectors; and recommending data products to the data buyer based on the similarity. The embodiments of this application can fully meet the scenario-based needs of data buyers.

Description

Technical Field

[0001] This application relates to the field of data technology, and in particular to a data product recommendation method, apparatus, device, storage medium, and product. Background Technology

[0002] In recent years, various data commodity operation and trading platforms have emerged one after another. Their main purpose is to promote the development of data elements and release the value of data, and to ensure the compliance of data resources.

[0003] Data commodity trading platforms typically serve two main stakeholders: data providers and data buyers. The platform usually retrieves and recommends data products from data providers based on keywords entered by the data buyer. For example, if an automotive company wants to identify customers with potential demand for new energy vehicles, it can input keywords such as travel habit data, car consumption data, and car purchase consultation data to obtain data products related to car use and consumption. However, relying solely on fragmented keywords entered by the data buyer is insufficient to comprehensively describe the specific application scenario's needs, resulting in the platform recommending data products that fail to effectively meet the data buyer's requirements. Summary of the Invention

[0004] This application provides a data product recommendation method, apparatus, device, storage medium, and product to solve the problem in the prior art that it is difficult to fully describe the demand information of specific application scenarios by only inputting keywords, resulting in the data products recommended by the platform failing to effectively meet the scenario needs of data buyers.

[0005] To achieve the above objectives, embodiments of this application provide a data product recommendation method, including: Obtain scenario-specific requirements information input by the data buyer; Extract keywords from the scenario requirement information to form a target keyword set; The target keyword set is vectorized to form the first set of vectors; The text information from multiple data providers is vectorized to obtain multiple second sets of vectors; Calculate the similarity between the first group of vectors and each of the second group of vectors, and recommend data products to the data buyer based on the similarity.

[0006] In one optional embodiment, the step of extracting keywords from the scenario requirement information to form a target keyword set includes: Extract several keywords from the scenario requirements information; Several derived keywords are obtained by further deriving from the aforementioned keywords; The aforementioned keywords and the aforementioned derived keywords are used as the target keyword set.

[0007] In one optional embodiment, the extraction of several keywords from the scenario requirement information includes: The parameters of the language representation model are optimized to obtain the target language representation model; Using the target language representation model, the keywords are extracted from the scenario requirement information.

[0008] In one optional embodiment, the derivation of the plurality of keywords to obtain a plurality of derived keywords includes: For each of the aforementioned keywords, corpora similar to the keyword are obtained from a preset corpus database as derived keywords of the keyword.

[0009] In an optional embodiment, calculating the similarity between the first set of vectors and each of the second set of vectors includes: Each similarity is obtained based on the cosine distance between the first group of vectors and each of the second group of vectors.

[0010] In one optional embodiment, the text information of the data provider includes at least one of the following: Company description; Data product description; Data service description.

[0011] To achieve the above objectives, embodiments of this application also provide a data product recommendation device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the data product recommendation method as described above.

[0012] To achieve the above objectives, embodiments of this application also provide a computer-readable storage medium, the computer-readable storage medium including a stored computer program; wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the data product recommendation method as described above.

[0013] To achieve the above objectives, embodiments of this application also provide a computer program product, including a computer program / instructions, which, when executed by a processor, implement the data product recommendation method as described above.

[0014] Compared with existing technologies, the data product recommendation method, apparatus, device, storage medium, and product provided in this application embodiment obtain scenario demand information input by a data buyer; extract keywords from the scenario demand information to form a target keyword set; vectorize the target keyword set to form a first set of vectors; vectorize text information from multiple data providers to obtain multiple second sets of vectors; calculate the similarity between the first set of vectors and each of the second set of vectors; and recommend data products to the data buyer based on the similarity. Therefore, this application embodiment transforms the existing retrieval mode of inputting scattered keywords into data product recommendation based on inputting complete scenario demand information. This effectively reduces the problems of demand information deviation and retrieval scope attenuation caused by data buyers independently extracting keywords, making the recommendation results more consistent with the data buyer's needs, improving the adaptability of the recommendation results to scenario requirements, and fully satisfying the scenario-based needs of the data buyer. Attached Figure Description

[0015] Figure 1 This is a flowchart of a data product recommendation method provided in an embodiment of this application; Figure 2 This is a schematic diagram of a keyword search interface interaction design. Figure 3 This is a schematic diagram of an interactive design for a demand retrieval interface provided in an embodiment of this application; Figure 4 This is a schematic diagram of a vectorization process provided in an embodiment of this application; Figure 5 This is a schematic diagram of another vectorization process provided in an embodiment of this application; Figure 6 This is a structural block diagram of a data product recommendation device provided in an embodiment of this application; Figure 7 This is a structural block diagram of a data product recommendation device provided in an embodiment of this application. Detailed Implementation

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0017] In the description of this application, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0018] In this application description, the terms "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0019] In this application description, the terms "first," "second," etc., are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus. The term "based on" means "at least partially based on." The term "according to" means "at least partially according to." The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments." The term "and / or" means at least one of the connected objects, such as A and / or B, indicating three cases: including only A, only B, and both A and B. Unless otherwise stated, the term "multiple" means two or more.

[0020] It is understood that before using the technical solutions disclosed in the various embodiments of this disclosure, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this disclosure in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0021] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose whether to provide personal information to the software or hardware, such as the electronic device, application, server, or storage medium performing the operations of this disclosed technical solution, based on the prompt message.

[0022] As an optional but non-limiting implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device. It is understood that the above notification and user authorization process is merely illustrative and does not constitute a limitation on the implementation of this disclosure; other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure.

[0023] It is understood that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0024] See Figure 1 , Figure 1 This is a flowchart of a data product recommendation method provided in an embodiment of this application. The data product recommendation method includes: S1. Obtain scenario requirement information input by the data buyer; This application's embodiments shift from the existing retrieval mode of inputting scattered keywords to data product recommendations based on inputting complete scenario-specific needs information. To support this shift, methods such as... Figure 2 The search interface interaction has been changed from the commonly used keyword search interface interaction design to something like... Figure 3 The requirements retrieval interface interaction design includes: prompts and input examples; the prompts ask for a textual description of the data application scenario requirements, and the input examples display the specific textual descriptions to help users understand and input the data. The textual descriptions of the data application scenario requirements entered here constitute the scenario requirement information, reflecting the user's actual needs for a specific application scenario.

[0025] Data buyers can directly input textual descriptions of their scenario requirements, eliminating the need to extract multiple keywords and perform multiple searches, as shown in Table 1. Table 1

[0026] S2. Extract keywords from the scenario requirement information to form a target keyword set; This application embodiment can utilize Natural Language Processing (NLP) technology to perform intent recognition on scene requirement information and extract keywords expressing the requirements to form a target keyword set, which can accurately express the semantic information in the scene requirement information.

[0027] For example, common language representation models such as FastText, BERT (Bidirectional Encoder Representations from Transformers), GPT (Generative Pre-trained Transformer), and T5 (Text-to-Text Transfer Transformer) can be used for keyword extraction. Considering factors such as model performance, deployment cost, and ease of use, this application's embodiments preferably use the BERT model for keyword extraction.

[0028] S3. Vectorize the target keyword set to form the first set of vectors; In order to quickly calculate the data provider and data product information corresponding to the target keyword set, this embodiment of the application needs to transform the keywords in the target keyword set from uncalculable and unstructured words into calculable and structured vector values, so as to complete the retrieval of massive information through similarity calculation, and achieve the goal of fast, accurate and comprehensive data element recommendation.

[0029] The vectorization technique for the target keyword set mainly adopts the word-vector model (Word2vec model). Word2vec is a deep learning model for natural language processing, whose main purpose is to represent words as vectors in a low-dimensional vector space. These vectors can capture the semantic and syntactic information of words, so that semantically similar words are close in distance within this vector space.

[0030] The Word2vec model includes two network structures: the CBOW (Continues Bag of Words) model and the Skip-gram model. Since the CBOW model predicts the center word after integrating the context words, it is not as good as the Skip-gram model in capturing the subtle semantic differences between words. Moreover, it is mainly suitable for tasks with low requirements for text semantics. Therefore, this application embodiment preferably uses the Skip-gram model structure to vectorize the target keyword set so as to capture the semantic and grammatical information of words well, thereby improving the performance of the system.

[0031] The target keyword set is vectorized using the Word2vec model, mapping each keyword to a high-dimensional real-number vector to form the first set of vectors. An example of vectorization for the target keyword set used to build a model of potential customers for new energy vehicle upgrades and replacements is shown below. Figure 4 .

[0032] S4. Vectorize the text information from multiple data providers to obtain multiple second sets of vectors; It is worth noting that data providers input relevant text information into the data commodity operation and trading platform for data buyers to search and view. This input information is mainly provided in the form of text content and is primarily described from the perspective of the data provider. This differs somewhat from the potential data application scenarios and needs of data buyers in various industries. If suitable keywords cannot be extracted, the data provider's services cannot be found. Therefore, this example vectorizes the text information of the data provider, representing it as a vector that can express the semantics of the text. This is beneficial for data buyers to search for and discover suitable data elements through similarity calculation.

[0033] This application embodiment can utilize word embedding to vectorize text information from multiple data providers, obtaining a second set of vectors for each data provider. Common word embedding methods include the One-Hot Model, Bag of Words Model, Term Frequency-Inverse Document Frequency (TF-IDF), N-gram model, Word2vec, and Document2vec.

[0034] Optionally, the text information of the data provider includes at least one of the following: company description, data product description, and data service description.

[0035] Since this embodiment requires vectorizing unstructured text descriptions such as enterprise descriptions, data product descriptions, and data service descriptions of data providers, the Doc2vec model, which considers both word and paragraph relationships, is preferred because it captures the semantic information of paragraphs, not just individual words. Furthermore, the Doc2vec model has a fast training speed, can handle large-scale text data, and allows for incremental updates, making it more flexible. Therefore, this embodiment preferably uses Doc2vec to vectorize the text information of data providers.

[0036] For example, for a data provider that offers car travel data and provides data products to companies that target car owners for marketing services, the process of vectorizing data based on the company description is as follows: Figure 5 The number of words segmented in the enterprise description (e.g.) Figure 5 The number of Chinese text segments is [batch_size, N], and the second set of vectors (such as...) Figure 5 The corresponding Chinese text vector is [batch_size, N, 4]; where batch_size represents the batch size, and N represents the number of word segments. S5. Calculate the similarity between the first group of vectors and each of the second group of vectors, and recommend data products to the data buyer based on the similarity.

[0037] It's worth noting that after vectorization, two sets of computable, structured vector values ​​are formed (i.e., the first set of vectors and the second set of vectors). To find data elements that closely match the scenario requirements, the similarity between the first set of vectors and each element in the second set of vectors is calculated. Based on this similarity, suitable data products are then selected for recommendation. For example, the n data products with the highest similarity are recommended to the data buyer, or data products are output in descending order of similarity and recommended to the data buyer; no specific restrictions are imposed here.

[0038] In one optional embodiment, the step of extracting keywords from the scenario requirement information to form a target keyword set includes: Extract several keywords from the scenario requirements information; Several derived keywords are obtained by further deriving from the aforementioned keywords; The aforementioned keywords and the aforementioned derived keywords are used as the target keyword set.

[0039] It is worth noting that, in order to retrieve data providers and data products that meet the needs of different scenarios from multiple dimensions, several keywords extracted from the scenario requirement information will be derived into multiple keywords related to the scenario requirements as derivative keywords, forming a broader and more comprehensive set of target keywords. This will enable more comprehensive data element recommendations and maximize the utilization value of data elements.

[0040] In one optional embodiment, the extraction of several keywords from the scenario requirement information includes: The parameters of the language representation model are optimized to obtain the target language representation model; Using the target language representation model, the keywords are extracted from the scenario requirement information.

[0041] In order to improve the accuracy of keyword extraction, this application embodiment optimizes the parameters of the language representation model to obtain a target language representation model for keyword extraction.

[0042] In one specific embodiment, the BERT model is selected, and the parameters of the BERT model are optimized using the Robustly Optimized BERT Approach (RoBERTa). This method mainly improves BERT performance through the following four methods: ① longer training steps, larger batch size, and more data; ② removing the NSP (NextSentence Prediction) training objective; ③ longer sequences; ④ changing the static mask to a dynamic mask.

[0043] The goal of training the BERT model by optimizing the RoBERTa parameters is to maximize the likelihood probability of the following function:

[0044] in, To mask the flag bit, serving as a summation and filtering function, if the first bit in the sequence... If a word is obscured, then =1, otherwise =0; For the position located at the The words in each position are obscured.

[0045] The BERT model, after parameter optimization and adjustment via RoBERTa, can extract keywords from the scenario requirements input by the data buyer, forming a target keyword set, as shown in Table 2 below: Table 2

[0046] In one optional embodiment, the derivation of the plurality of keywords to obtain a plurality of derived keywords includes: For each of the aforementioned keywords, corpora similar to the keyword are obtained from a preset corpus database as derived keywords of the keyword.

[0047] Optionally, the pre-defined corpus database includes at least one of WordNet and ChineseSemanticKB. WordNet is an English word-derived corpus database that organizes English words into a series of synsets, each synset representing a concept and containing a set of near-synonyms for that concept, as well as their relationships with other concepts.

[0048] ChineseSemanticKB is a Chinese word derivation corpus database. The ChineseSemanticKB corpus database contains 12 categories and millions of commonly used semantic dictionaries, including a 340,000-word abstract lexicon, a 340,000-word antonym lexicon, and a 430,000-word synonym lexicon. Through these lexicons, word derivation operations can be performed on words, including abstraction, antonymization, synonymization, and abbreviation.

[0049] In order to support the needs of data buyers in scenarios using both Chinese and English, this application embodiment will use both WordNet and ChineseSemanticKB to derive the target keyword set. By deploying a pre-set corpus in the private environment of the data commodity operation and trading platform, the cost of using the corpus can be saved, the speed of accessing the corpus can be improved, and the corpus data can be perfected.

[0050] Examples of generating derived keywords using a pre-defined corpus are shown in Table 3: Table 3

[0051] In an optional embodiment, calculating the similarity between the first set of vectors and each of the second set of vectors includes: Each similarity is obtained based on the cosine distance between the first group of vectors and each of the second group of vectors.

[0052] It is worth noting that there are three main methods for calculating similarity based on vectors: cosine distance, Euclidean distance, and Jaccard distance. Euclidean distance primarily calculates the absolute difference in individual numerical features, used for analyses that require differences to be reflected in the magnitude of dimensional values. Jaccard distance, on the other hand, does not consider the magnitude of potential values ​​in vectors, simply treating them as 0 and 1. Therefore, this embodiment uses the cosine distance algorithm. Cosine distance distinguishes differences more from a directional perspective and is not sensitive to absolute numerical values, while also correcting the potential problem of inconsistent measurement standards between vector values.

[0053] The formula for calculating cosine distance is as follows:

[0054] in, This represents the cosine distance between the first set of vectors x and the second set of vectors y. This represents the components of the first vector x. Let represent the components of the second group of vectors y, and n represent the number of components of the first group of vectors x and the second group of vectors y.

[0055] The cosine distance between the first and second sets of vectors is used as the similarity between them. When the cosine distance equals 1, it means that the two sets of information are completely duplicated; when the cosine distance is close to 1, it means that the second set of vectors is similar to the first set of vectors, which can be considered to meet the retrieval requirements and can be recommended to the data buyer; the smaller the cosine distance, the less relevant the second set of vectors is to the first set of vectors.

[0056] For example, in response to the needs of car manufacturers for identifying potential customers for new energy vehicle replacement and upgrades, the data is output in descending order of cosine distance. The larger the cosine distance value, the higher the match with the keywords. Although the matching degree of the following data elements is relatively lower, they are also output as recommendations, giving data buyers more dimensions of data to choose from. The final output results and order are shown in Table 4. Table 4

[0057] See Figure 6 , Figure 6 This is a structural block diagram of a data product recommendation device 10 provided in an embodiment of this application. The data product recommendation device 10 includes: The acquisition module 11 is used to acquire the scenario requirement information input by the data buyer; Extraction module 12 is used to extract keywords from the scenario requirement information to form a target keyword set; The first vectorization module 13 is used to perform vectorization processing on the target keyword set to form a first set of vectors; The second vectorization module 14 is used to vectorize the text information from multiple data providers to obtain multiple second sets of vectors. Recommendation module 15 is used to calculate the similarity between the first group of vectors and each of the second group of vectors, and recommend data products to the data buyer based on the similarity.

[0058] Optionally, the step of extracting keywords from the scenario requirement information to form a target keyword set includes: Extract several keywords from the scenario requirements information; Several derived keywords are obtained by further deriving from the aforementioned keywords; The aforementioned keywords and the aforementioned derived keywords are used as the target keyword set.

[0059] Optionally, the extraction of several keywords from the scenario requirement information includes: The parameters of the language representation model are optimized to obtain the target language representation model; Using the target language representation model, the keywords are extracted from the scenario requirement information.

[0060] Optionally, the process of deriving from the aforementioned keywords to obtain a number of derived keywords includes: For each of the aforementioned keywords, corpora similar to the keyword are obtained from a preset corpus database as derived keywords of the keyword.

[0061] Optionally, calculating the similarity between the first set of vectors and each of the second set of vectors includes: Each similarity is obtained based on the cosine distance between the first group of vectors and each of the second group of vectors.

[0062] Optionally, the text information of the data provider includes at least one of the following: Company description; Data product description; Data service description.

[0063] It is worth noting that the working process of each module in the data product recommendation device 10 described in this application embodiment can refer to the working process of the data product recommendation method described in the above embodiment and achieve the same beneficial effects, and will not be repeated here.

[0064] Furthermore, this application also provides a computer-readable storage medium, which includes a stored computer program; wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the data product recommendation method as described in any of the above embodiments.

[0065] Furthermore, this application also provides a computer program product, including a computer program / instructions, which, when executed by a processor, implements the data product recommendation method as described in any of the above embodiments.

[0066] See Figure 7 , Figure 7 This is a structural block diagram of a data product recommendation device 20 provided in an embodiment of this application. The data product recommendation device 20 includes: a processor 21, a memory 22, and a computer program stored in the memory 22 and executable on the processor 21. When the processor 21 executes the computer program, it implements the steps in the above-described data product recommendation method embodiments. Alternatively, when the processor 21 executes the computer program, it implements the functions of each module / unit in the above-described device embodiments.

[0067] For example, the computer program may be divided into one or more modules / units, which are stored in the memory 22 and executed by the processor 21 to complete this application. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the data product recommendation device 20.

[0068] The data product recommendation device 20 may include, but is not limited to, a processor 21 and a memory 22. Those skilled in the art will understand that the schematic diagram is merely an example of the data product recommendation device 20 and does not constitute a limitation on the data product recommendation device 20. It may include more or fewer components than illustrated, or combine certain components, or different components. For example, the data product recommendation device 20 may also include input / output devices, network access devices, buses, etc.

[0069] The processor 21 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor 21 is the control center of the data product recommendation device 20, connecting all parts of the data product recommendation device 20 via various interfaces and lines.

[0070] The memory 22 can be used to store the computer programs and / or modules. The processor 21 implements various functions of the data product recommendation device 20 by running or executing the computer programs and / or modules stored in the memory 22 and calling the data stored in the memory 22. The memory 22 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory 22 may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0071] If the modules / units integrated in the data product recommendation device 20 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by the processor 21, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0072] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided in this application, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0073] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications are also considered to be within the scope of protection of this application.

Claims

1. A data product recommendation method, characterized in that, include: Obtain scenario-specific requirements information input by the data buyer; Extract keywords from the scenario requirement information to form a target keyword set; The target keyword set is vectorized to form the first set of vectors; The text information from multiple data providers is vectorized to obtain multiple second sets of vectors; Calculate the similarity between the first group of vectors and each of the second group of vectors, and recommend data products to the data buyer based on the similarity.

2. The data product recommendation method as described in claim 1, characterized in that, The extraction of keywords from the scenario requirement information to form a target keyword set includes: Extract several keywords from the scenario requirements information; Several derived keywords are obtained by further deriving from the aforementioned keywords; The aforementioned keywords and the aforementioned derived keywords are used as the target keyword set.

3. The data product recommendation method as described in claim 2, characterized in that, The extraction of several keywords from the scenario requirement information includes: The parameters of the language representation model are optimized to obtain the target language representation model; Using the target language representation model, the keywords are extracted from the scenario requirement information.

4. The data product recommendation method as described in claim 2, characterized in that, The process of deriving from the aforementioned keywords yields several derived keywords, including: For each of the aforementioned keywords, corpora similar to the keyword are obtained from a preset corpus database as derived keywords of the keyword.

5. The data product recommendation method as described in claim 1, characterized in that, The calculation of the similarity between the first group of vectors and each of the second group of vectors includes: Each similarity is obtained based on the cosine distance between the first group of vectors and each of the second group of vectors.

6. The data product recommendation method as described in claim 1, characterized in that, The text information from the data provider includes at least one of the following: Company description; Data product description; Data service description.

7. A data product recommendation device, characterized in that, include: The acquisition module is used to acquire scenario requirement information input by the data buyer; The extraction module is used to extract keywords from the scenario requirement information to form a target keyword set; The first vectorization module is used to perform vectorization processing on the target keyword set to form a first set of vectors; The second vectorization module is used to vectorize the text information from multiple data providers to obtain multiple second sets of vectors. The recommendation module is used to calculate the similarity between the first group of vectors and each of the second group of vectors, and recommend data products to the data buyer based on the similarity.

8. A data product recommendation device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the data product recommendation method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program; wherein, when the computer program is executed, it controls the device on which the computer-readable storage medium is located to perform the data product recommendation method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, Includes a computer program / instruction that, when executed by a processor, implements the data product recommendation method as described in any one of claims 1 to 6.

Images