Method and device for generating title

Abstract

The present invention discloses a method and apparatus for generating titles, relating to the field of computer technology. A specific implementation of the method includes extracting search keywords from historical search data; clustering the search keywords to generate a keyword dataset; analyzing the keyword dataset to obtain core keywords and their core weights; and generating titles based on the core weights of the core keywords. This implementation can improve the hit conversion rate and user experience; the generated titles accurately capture the product features of interest to users and are consistent with the product content.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and device for generating a title. Background Art

[0002] With the popularization of the Internet and the development of computer technology, the Internet has long been integrated into people's lives and brought convenience to life. At present, people have become accustomed to online life, using it to purchase goods, check news, search for papers, and so on.

[0003] Typically, users search and browse related products using keywords. When displaying products (such as commodities, news, or papers), various websites and platforms only display the product title to reduce page size or facilitate user browsing. The title is a brief introduction generated based on the product's content or features. Currently, there are two main methods for generating titles:

[0004] 1. Extraction: Extract the summary sentences and words that generate the title from the original text of the product, compress the summary sentences and words, and then generate the title;

[0005] 2. Generative formula: Generate summary sentences and words of the title, perform text compression on the summary sentences and words, and thus generate the title. The summary sentences and words can be generated freely and are not required to be extracted from the original text.

[0006] At present, in order to improve the search hit rate of products, increase the purchase volume, news reading volume or paper reading volume, etc., most website platforms will make the product titles cover multiple categories of keywords to the maximum extent and increase the text length of the titles.

[0007] In the process of implementing the present invention, the inventors discovered that the prior art has at least the following problems:

[0008] The titles generated by existing technologies have problems such as being inaccurate, too long or redundant. Keyword searches can easily result in too many products being found, or some of the products found not matching the keywords. This results in a low conversion rate and poor user experience. Summary of the Invention

[0009] In view of this, an embodiment of the present invention provides a method and device for generating titles, which can improve the hit conversion rate and user experience; the generated titles accurately include the product features that users are concerned about and are consistent with the content of the product.

[0010] To achieve the above objective, according to one aspect of an embodiment of the present invention, a method for generating a title is provided.

[0011] A method for generating a title in an embodiment of the present invention includes: extracting search keywords from historical search data; clustering the search keywords to obtain a keyword data set; analyzing the keyword data set to obtain core keywords and core weights of the core keywords; and generating a title based on the core weights of the core keywords.

[0012] Optionally, analyzing the keyword data set to obtain core keywords and core weights of the core keywords includes: calculating the average value of the keyword data set as the core keyword; assigning initial weights to the core keywords; using a bidirectional neural network to calculate the similarity values ​​between the keyword data sets; adjusting the initial weights based on the similarity values ​​to obtain adjusted weights, and selecting the largest adjusted weight as the core weight.

[0013] Optionally, adjusting the initial weight based on the similarity value to obtain the adjusted weight includes: adjusting the initial weight using a weight processing formula to obtain an adjusted weight, and selecting the largest adjusted weight as the core weight; wherein the weight processing formula is: W=v+α(xv); W is the adjusted weight, v is the initial weight, x is the adjusted independent variable, and α is the similarity value.

[0014] Optionally, clustering the search keywords includes clustering the search keywords using a K-means algorithm, a K-medoid algorithm, a density-based clustering algorithm, or a Gaussian mixture model.

[0015] Optionally, clustering the search keywords to obtain a keyword data set includes: randomly selecting K search keywords as cluster centroid points; dividing the search keywords with the same category as the cluster centroid points into a keyword cluster; calculating the average value of the keyword cluster as a new cluster centroid point; re-dividing the search keywords with the same category as the new cluster centroid point into a keyword cluster; and taking the keyword cluster whose cluster centroid points no longer change or the number of divisions reaches a preset value as the keyword data set.

[0016] Optionally, extracting the search keyword from the historical search data includes: clustering the historical search data to obtain a historical data set; and calculating an average value of the historical data set as the search keyword.

[0017] Optionally, the method further includes: extracting real-time keywords based on real-time search data; clustering the real-time keywords to obtain a real-time data set; analyzing the real-time data set to obtain verification keywords and verification weights of the verification keywords; and verifying or adjusting the title based on the verification weights of the verification keywords.

[0018] To achieve the above objective, according to another aspect of an embodiment of the present invention, a device for generating a title is provided.

[0019] A device for generating a title in an embodiment of the present invention includes: an extraction module for extracting search keywords from historical search data; a clustering module for clustering the search keywords to obtain a keyword data set; an analysis module for analyzing the keyword data set to obtain core keywords and core weights of the core keywords; and a generation module for generating a title based on the core weights of the core keywords.

[0020] Optionally, the analysis module is also used to: calculate the average value of the keyword data set as the core keyword; assign initial weights to the core keywords; use a bidirectional neural network to calculate the similarity values ​​between the keyword data sets; adjust the initial weights based on the similarity values ​​to obtain adjusted weights, and select the largest adjusted weight as the core weight.

[0021] Optionally, the analysis module is further used to: adjust the initial weight using a weight processing formula to obtain an adjusted weight, and select the largest adjusted weight as the core weight; wherein the weight processing formula is: W = v + α (xv); W is the adjusted weight, v is the initial weight, x is the adjusted independent variable, and α is the similarity value.

[0022] Optionally, the clustering module is further configured to cluster the search keywords using a K-means algorithm, a K-center point algorithm, a density-based clustering algorithm, or a Gaussian mixture model.

[0023] Optionally, the clustering module is also used to: randomly select K search keywords as cluster centroid points; divide the search keywords with the same category as the cluster centroid point into a keyword cluster; calculate the average value of the keyword cluster as the new cluster centroid point; re-divide the search keywords with the same category as the new cluster centroid point into a keyword cluster; and take the keyword cluster whose cluster centroid point no longer changes or the number of divisions reaches a preset value as the keyword data set.

[0024] Optionally, the extraction module is further configured to: cluster the historical search data to obtain a historical data set; and calculate an average value of the historical data set as a search keyword.

[0025] Optionally, the device further includes: a verification module, configured to extract real-time keywords based on real-time search data; cluster the real-time keywords to obtain a real-time data set; analyze the real-time data set to obtain verification keywords and verification weights of the verification keywords; and verify or adjust the title based on the verification weights of the verification keywords.

[0026] To achieve the above objective, according to another aspect of an embodiment of the present invention, an electronic device for generating a title is provided.

[0027] An electronic device for generating a title according to an embodiment of the present invention includes: one or more processors; a storage device for storing one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors implement a method for generating a title according to an embodiment of the present invention.

[0028] To achieve the above objective, according to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided.

[0029] A computer-readable storage medium according to an embodiment of the present invention stores a computer program, which, when executed by a processor, implements a method for generating a title according to an embodiment of the present invention.

[0030] One embodiment of the above invention has the following advantages or beneficial effects: because it adopts the method of extracting search keywords from historical search data; clustering the search keywords to obtain a keyword data set; analyzing the keyword data set to obtain core keywords and core weights of core keywords; and generating titles based on the core weights of core keywords, and generating titles according to real user operation behaviors, it overcomes the technical problems of inaccurate titles, too long or redundant texts, low hit conversion rates, and poor user experience, thereby achieving the goal of improving hit conversion rates and user experience; the generated titles accurately contain the product features that users are concerned about, and are consistent with the content of the products.

[0031] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The accompanying drawings are provided for a better understanding of the present invention and are not intended to limit the present invention.

[0033] Figure 1 is a schematic diagram of the main steps of a method for generating a title according to an embodiment of the present invention;

[0034] Figure 2 is a schematic diagram of the main process of a method for generating a title according to a reference embodiment of the present invention;

[0035] Figure 3 is a schematic diagram of the main process of clustering search keywords according to a reference embodiment of the present invention;

[0036] Figure 4 is a schematic diagram of main modules of an apparatus for generating a title according to an embodiment of the present invention;

[0037] Figure 5is an exemplary system architecture diagram in which embodiments of the present invention may be applied;

[0038] Figure 6 It is a schematic diagram of the structure of a computer system of a terminal device or server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0039] The following description of exemplary embodiments of the present invention is made in conjunction with the accompanying drawings, in which various details of the embodiments of the present invention are included to facilitate understanding. These details should be considered as merely exemplary. Therefore, it should be appreciated by those skilled in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0040] It should be pointed out that, in the absence of conflict, the embodiments of the present invention and the technical features therein may be combined with each other.

[0041] Figure 1 FIG. 4 is a schematic diagram of main steps of a method for generating a title according to an embodiment of the present invention.

[0042] like Figure 1 As shown, the method for generating a title in an embodiment of the present invention mainly includes the following steps:

[0043] Step S101: extracting search keywords from historical search data.

[0044] In existing technologies, to maximize keyword coverage, titles are often long and redundant. This can result in a high proportion of searched products not matching the keywords, leading to a poor user experience. Furthermore, due to excessively long or redundant titles, keyword searches can easily return an excessive number of products, reducing conversion rates.

[0045] Typically, a product title significantly impacts its conversion rate. Imprecise titles, excessive length, or overly brief titles can all reduce this rate. The conversion rate refers to the probability that a user will find a product through a keyword search and then proceed with actions such as purchasing, viewing, or adding it to a favorite. To improve user experience and conversion rates, embodiments of the present invention propose a method for generating titles based on actual user behavior.

[0046] In order to enable users to find the expected products quickly and accurately, that is, to ensure that the generated title contains the product features that the user is interested in and is consistent with the content of the product, an embodiment of the present invention extracts search keywords from the historical search data corresponding to the product. The historical search data is the content (keywords) searched by the user. After the user finds the product by searching for a certain keyword, the user purchases, views or collects the product. Then, the keyword is the historical search data of the product. For example, after the user searches for "computer paper", he collects "A paper" in the search results. Then, "computer paper" can be used as the historical search data of the product "A paper". In addition, for extracting search keywords from historical search data, the keywords in the historical search data can be counted to obtain the search frequency of each keyword, and several keywords with high search frequency can be selected as search keywords; search keywords can also be extracted from historical search data based on certain algorithm models, such as the artificial intelligence learning system (TensorFlow) framework. TensorFlow is a framework that transfers complex data structures to artificial intelligence neural networks for analysis and processing.

[0047] In an embodiment of the present invention, step S101 may be implemented in the following manner: clustering the historical search data to obtain a historical data set; and calculating an average value of the historical data set as the search keyword.

[0048] Clustering refers to identifying inherent rules within data and dividing the data into several categories according to these rules. Clustering historical search data involves categorizing it. The historical search data within each cluster constitutes a historical dataset, with high similarity values ​​between historical search data within the same dataset and low similarity values ​​between historical search data from different datasets. The average value of a historical dataset refers to the historical search data within the dataset with the shortest distance from the cluster center, i.e., the historical search data closest to the cluster center. The distance between historical search data and the cluster center can be measured using Minkowski distance, Manhattan distance, or Chebyshev distance, among other metrics.

[0049] Step S102: clustering the search keywords to obtain a keyword data set.

[0050] The search keywords extracted in step S101 may be classified, and search keywords of the same category may be placed in the same keyword data set.

[0051] In an embodiment of the present invention, clustering of search keywords may be performed in the following manners: clustering of search keywords using a K-means algorithm, a K-medoid algorithm, a density-based clustering algorithm, or a Gaussian mixture model.

[0052] Among them, the K-means algorithm (k-Means) divides n search keywords into K clusters so that the obtained clusters satisfy: the similarity values ​​of search keywords in the same cluster are high; while the similarity values ​​of search keywords in different clusters are small. The K-medoids algorithm (K-MEDOIDS) selects a search keyword as the center point, identifies the cluster through the center point, and assigns the remaining search keywords to the corresponding clusters according to the category of the search keyword. The density-based clustering algorithm (DBSCAN) finds the largest set of density-connected search keywords. The Gaussian mixture model (GMM) gives the probability that each search keyword belongs to each cluster and assigns the search keyword to the cluster corresponding to the highest probability.

[0053] In an embodiment of the present invention, step S102 can be implemented in the following manner: randomly selecting K search keywords as cluster centroid points; dividing the search keywords with the same category as the cluster centroid point into a keyword cluster; calculating the average value of the keyword cluster as the new cluster centroid point; re-dividing the search keywords with the same category as the new cluster centroid point into a keyword cluster; and taking the keyword cluster whose cluster centroid point no longer changes or the number of iterations reaches a preset value as the keyword data set.

[0054] Different search keywords have different cluster centroids for the same product, and the keyword clusters obtained by selecting different cluster centroids may also be different. Therefore, the cluster centroids can be reselected based on the results of the previous clustering and clustering can be performed again until the keyword clusters are relatively stable (i.e., the cluster centroids no longer change) or the number of repeated clustering (i.e., the number of iterations) reaches a preset value. The average value of a keyword cluster refers to the search keyword with the smallest distance from the cluster center in the keyword cluster, that is, the search keyword closest to the cluster center. Iteration refers to the process of repeatedly executing a series of calculation steps to sequentially calculate the subsequent quantities from the previous quantities. The above method can be used to calculate the final cluster content (i.e., keyword data set).

[0055] Step S103: Analyze the keyword data set to obtain core keywords and core weights of the core keywords.

[0056] Core keywords are the data that best represents the category of a keyword dataset. By analyzing a keyword dataset, we can identify the search keyword within a category that best reflects user expectations. This search keyword is the core keyword. The core weight reflects the contribution of the core keyword to user decisions to purchase, view, or save an item, indicating the degree of influence the core keyword has on the user's decision.

[0057] In an embodiment of the present invention, step S103 can be implemented in the following manner: calculating the average value of the keyword data set as the core keyword; assigning initial weights to the core keywords; using a bidirectional neural network to calculate the similarity values ​​between the keyword data sets; adjusting the initial weights based on the similarity values ​​to obtain adjusted weights, and selecting the largest adjusted weight as the core weight.

[0058] The average value of the keyword data set refers to the search keyword with the smallest distance from the cluster center in the keyword data set, that is, the search keyword closest to the cluster center. In step S102, clustering is performed for different search keywords, but the relationship between the keyword data sets (that is, the relationship between the search keywords) is not clear. Therefore, a bidirectional neural network can be used to analyze the relationship between the keyword data sets, and the core weight of the core keyword can also be calculated using the bidirectional neural network. Among them, the initial weight assigned to the core keyword can be the proportion of the core keyword in the keyword data set, or the degree of influence of the core keyword category on the user's decision. The similarity value is a quantification of the relationship between any two keyword data sets, and each keyword data set may correspond to multiple similarity values. Therefore, adjusting the initial weight may result in multiple adjusted weights, and the largest adjusted weight can be selected as the core weight of the core keyword.

[0059] In an embodiment of the present invention, the initial weights are adjusted based on the similarity values ​​to obtain the adjusted weights in the following manner: the initial weights are adjusted using a weight processing formula to obtain the adjusted weights, and the largest adjusted weight is selected as the core weight. The weight processing formula is: W = v + α(xv); W is the adjusted weight, v is the initial weight, x is the independent variable to be adjusted, and α is the similarity value. The independent variable to be adjusted can be determined based on actual needs or past experience. The similarity values ​​can be used to roughly determine the connections between keyword data sets, and the size of the similarity values ​​can be used to determine which core keywords' initial weights need to be increased or decreased.

[0060] Step S104: Generate a title based on the core weight of the core keyword.

[0061] After obtaining the core weight of the core keyword, you can select the core keywords of the preset data with larger core weights to form the title of the product, or you can select the core keywords with core weights exceeding the preset value to form the title of the product, or you can select several core keywords with larger core weights to generate the title of the product based on a certain algorithm, and so on.

[0062] In an embodiment of the present invention, the method for generating a title may further include: extracting real-time keywords based on real-time search data; clustering the real-time keywords to obtain a real-time data set; analyzing the real-time data set to obtain verification keywords and verification weights of the verification keywords; and verifying or adjusting the title based on the verification weights of the verification keywords.

[0063] Steps S101-S104 generate titles based on historical search data, which is a prediction based on user operation behavior. The generated title is also a prediction. Therefore, real-time search data can also be collected. Based on a certain amount of real-time search data or real-time search data within a certain period of time, the generated title can be verified or adjusted. If the adjustment keyword and the core keyword, or the adjustment weight of the adjustment keyword and the core weight of the core keyword are the same, then the generated title basically meets the user's expectations and does not need to be adjusted. If the adjustment keyword and the core keyword, or the adjustment weight of the adjustment keyword and the core weight of the core keyword are different, then the title can be regenerated, the word order of the title can be adjusted, or the words in the title can be replaced. The analysis and processing process of real-time search data can adopt the same method as steps S101-S104.

[0064] It can be seen from the method for generating titles according to an embodiment of the present invention that the method extracts search keywords from historical search data; clusters the search keywords to obtain a keyword data set; analyzes the keyword data set to obtain core keywords and core weights of the core keywords; and generates titles based on the core weights of the core keywords. This technical method overcomes the technical problems of inaccurate titles, long or redundant texts, low hit conversion rates, and poor user experience, thereby achieving the goal of improving the hit conversion rate and user experience. The generated titles accurately include the product features that users are concerned about and are consistent with the content of the product.

[0065] Figure 2 It is a schematic diagram of the main process of a method for generating a title according to a reference embodiment of the present invention.

[0066] like Figure 2 As shown, the method for generating a title in the embodiment of the present invention can be implemented according to the following process:

[0067] Step S201: Obtain historical search data for a product:

[0068] Historical search data refers to the keywords that users searched for. After a user finds a product by searching for a keyword, and then purchases, views, or collects the product, the keyword becomes the historical search data for the product.

[0069] Step S202: Extract search keywords from historical search data of the product:

[0070] The historical search data can be clustered, and the historical search data in each cluster can be combined into a historical data set. The average value of the historical data set is calculated, and the average value of the historical data set is the search keyword.

[0071] Step S203: Cluster the search keywords to obtain a keyword data set:

[0072] The search keywords extracted in step S202 can be classified, and search keywords of the same category can be placed in the same keyword data set. It should be noted that clustering of search keywords can be achieved using a clustering algorithm such as the K-means algorithm, the K-medoid algorithm, a density-based clustering algorithm, or a Gaussian mixture model.

[0073] Step S204: Analyze the keyword dataset:

[0074] Analyzing keyword datasets yields core keywords and their core weights. The core keywords are calculated by calculating the average value of the keyword dataset. Using a bidirectional neural network to analyze the relationships between keyword datasets, similarity values ​​between them are obtained. Initial weights are adjusted using a weight processing formula to obtain adjusted weights, and the largest adjusted weight is selected as the core weight. The weight processing formula is: W = v + α(xv), where W is the adjusted weight, v is the initial weight, x is the adjusted independent variable, and α is the similarity value.

[0075] Step S205: Generate a title for the product based on the core weight of the core keyword.

[0076] After obtaining the core weight of the core keyword in step S204, you can select the core keywords of the preset data with larger core weights to form the title of the product, or you can select the core keywords with core weights exceeding the preset value to form the title of the product, or you can select several core keywords with larger core weights to generate the title of the product based on a certain algorithm, and so on.

[0077] Step S206: Verify or adjust the generated title based on the real-time search data:

[0078] Based on a certain amount of real-time search data or real-time search data within a certain period, the generated title can be verified or adjusted. If the adjusted keywords and the core keywords, or the adjusted weights of the adjusted keywords and the core weights of the core keywords are the same, then the generated title basically meets the user's expectations and does not need to be adjusted. If the adjusted keywords and the core keywords, or the adjusted weights of the adjusted keywords and the core weights of the core keywords are different, then the title can be regenerated, the word order of the title can be adjusted, or the words in the title can be replaced. It should be noted that the analysis and processing process for real-time search data can adopt the same method as steps S202 to S205.

[0079] Figure 3 1 is a schematic diagram of the main process of clustering search keywords according to a reference embodiment of the present invention.

[0080] like Figure 3 As shown, clustering search keywords in the method for generating titles in the embodiment of the present invention can be implemented according to the following process:

[0081] Step S301: Randomly select K search keywords as cluster centroids:

[0082] K represents the number of pre-specified keyword clusters.

[0083] Step S302: Divide the search keywords of the same category as the cluster centroid into a keyword cluster, and obtain K keyword clusters:

[0084] For each search keyword, use the following formula to calculate the keyword cluster it should belong to:

[0085] Among them, c (i) Represents a keyword cluster with the smallest distance between the cluster centroid and the search keyword among the K keyword clusters. represents the set of all independent variables j that minimize the distance between the cluster centroid and the search keyword, x (i) Represents a search keyword, μ j Indicates the search keyword (x (i) ) belongs to the cluster centroid point of the keyword cluster.

[0086] Step S303: Calculate the average value of the keyword cluster as the new cluster centroid:

[0087] For each keyword cluster, use the following formula to recalculate the cluster centroid of the keyword cluster to obtain the new cluster centroid:

[0088] Among them, μ j Indicates the search keyword (x(i) ) belongs to the cluster centroid point of the keyword cluster, m represents the number of keyword clusters in the keyword cluster, c (i) represents the keyword cluster with the smallest distance between the cluster centroid and the search keyword among the K keyword clusters, x (i) Represents a search keyword.

[0089] Step S304: Re-divide the search keywords of the same category as the new cluster centroid point into a keyword cluster.

[0090] Step S305: Determine whether the new cluster centroid point in step S303 is the same as the cluster centroid point in step S301, or whether the number of iterations reaches the preset value; if they are the same or have reached the preset value, execute step S306; if they are different and have not reached the preset value, execute step S303.

[0091] Step S306: Cluster the keywords in step S304 as a keyword dataset:

[0092] When the cluster centroid no longer changes or the number of iterations reaches the preset value, it means that the keyword clustering has become relatively stable. At this time, the keyword clustering can be used as a keyword dataset.

[0093] Figure 4 FIG. 4 is a schematic diagram of main modules of an apparatus for generating a title according to an embodiment of the present invention.

[0094] like Figure 4 As shown, the apparatus 200 for generating titles according to an embodiment of the present invention includes: an extraction module 401 , a clustering module 402 , an analysis module 403 and a generation module 404 .

[0095] in,

[0096] Extraction module 401, used to extract search keywords from historical search data;

[0097] Clustering module 402, configured to cluster the search keywords to obtain a keyword data set;

[0098] An analysis module 403 is configured to analyze the keyword data set to obtain core keywords and core weights of the core keywords;

[0099] The generating module 404 is configured to generate a title based on the core weight of the core keyword.

[0100] In an embodiment of the present invention, the analysis module 403 can also be used to: calculate the average value of the keyword data set as the core keyword; assign initial weights to the core keywords; use a bidirectional neural network to calculate the similarity values ​​between the keyword data sets; adjust the initial weights based on the similarity values ​​to obtain adjusted weights, and select the largest adjusted weight as the core weight.

[0101] In an embodiment of the present invention, the analysis module 403 can be further used to: adjust the initial weight using a weight processing formula to obtain an adjusted weight, and select the largest adjusted weight as the core weight; wherein the weight processing formula is: W = v + α (xv); W is the adjusted weight, v is the initial weight, x is the adjusted independent variable, and α is the similarity value.

[0102] In the embodiment of the present invention, the clustering module 402 may also be configured to cluster the search keywords using a K-means algorithm, a K-medoid algorithm, a density-based clustering algorithm, or a Gaussian mixture model.

[0103] In an embodiment of the present invention, the clustering module 402 can also be used to: randomly select K search keywords as cluster centroid points; divide the search keywords with the same category as the cluster centroid point into a keyword cluster; calculate the average value of the keyword cluster as a new cluster centroid point; re-divide the search keywords with the same category as the new cluster centroid point into a keyword cluster; and use the keyword cluster whose cluster centroid point no longer changes or the number of divisions reaches a preset value as a keyword data set.

[0104] In the embodiment of the present invention, the extraction module 401 may also be used to: cluster the historical search data to obtain a historical data set; and calculate an average value of the historical data set as a search keyword.

[0105] In addition, the device may also include: a verification module (not shown in the figure), which is used to extract real-time keywords based on real-time search data; cluster the real-time keywords to obtain a real-time data set; analyze the real-time data set to obtain verification keywords and verification weights of the verification keywords; and verify or adjust the title based on the verification weights of the verification keywords.

[0106] It can be seen from the device for generating titles according to an embodiment of the present invention that, because it adopts the technical means of extracting search keywords from historical search data; clustering the search keywords to obtain a keyword data set; analyzing the keyword data set to obtain core keywords and core weights of the core keywords; and generating titles based on the core weights of the core keywords, titles are generated according to real user operation behaviors, thereby overcoming the technical problems of inaccurate titles, long or redundant texts, low hit conversion rates, and poor user experience, thereby achieving the technical effect of improving the hit conversion rate and user experience; the generated titles accurately include the product features that users are concerned about and are consistent with the content of the product.

[0107] Figure 5 An exemplary system architecture 500 is shown to which the method for generating a title or the apparatus for generating a title according to an embodiment of the present invention can be applied.

[0108] like Figure 5 As shown, system architecture 500 may include terminal devices 501, 502, 503, a network 504, and a server 505. Network 504 is used to provide a medium for communication links between terminal devices 501, 502, 503 and server 505. Network 504 may include various connection types, such as wired or wireless communication links or fiber optic cables.

[0109] Users can use terminal devices 501, 502, and 503 to interact with server 505 via network 504 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 501, 502, and 503, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0110] The terminal devices 501 , 502 , and 503 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0111] Server 505 may be a server that provides various services, such as a backend management server that supports shopping websites browsed by users using terminal devices 501, 502, and 503. The backend management server may analyze and process received data such as product information query requests, and feed back the processing results (e.g., target push information, product information) to the terminal device.

[0112] It should be noted that the method for generating a title provided in the embodiment of the present invention is generally executed by the server 505 , and accordingly, the device for generating a title is generally set in the server 505 .

[0113] It should be understood that Figure 5The number of terminal devices, networks and servers in the embodiment is merely illustrative. Any number of terminal devices, networks and servers may be provided as required.

[0114] Reference below Figure 6 , which shows a schematic structural diagram of a computer system 600 of a terminal device suitable for implementing an embodiment of the present invention. Figure 6 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0115] like Figure 6 As shown, the computer system 600 includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage unit 608 into a random access memory (RAM) 603. Various programs and data required for the operation of the system 600 are also stored in the RAM 603. The CPU 601, ROM 602, and RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0116] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, and the like; an output section 607 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage section 608 including a hard disk; and a communication section 609 including a network interface card such as a LAN card or a modem. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 610 as needed, so that computer programs read therefrom can be installed into the storage section 608 as needed.

[0117] In particular, according to the embodiments disclosed in the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 609, and / or installed from a removable medium 611. When the computer program is executed by the central processing unit (CPU) 601, the above-mentioned functions defined in the system of the present invention are performed.

[0118] It should be noted that the computer-readable medium described in the present invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media can include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. This propagated data signal can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. Program code embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wireline, optical fiber cable, RF, or any suitable combination thereof.

[0119] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the above-mentioned module, program segment, or a part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0120] The modules described in the embodiments of the present invention may be implemented in software or hardware. The modules described may also be provided in a processor. For example, they may be described as comprising an extraction module, a clustering module, an analysis module, and a generation module. The names of these modules do not, in some cases, limit the modules themselves. For example, the extraction module may also be described as a "module for extracting search keywords from historical search data."

[0121] As another aspect, the present invention further provides a computer-readable medium, which may be included in the device described in the above embodiment, or may exist independently and not be incorporated into the device. The computer-readable medium carries one or more programs. When executed by the device, the device includes the following steps: Step S101: extracting search keywords from historical search data; Step S102: clustering the search keywords to obtain a keyword dataset; Step S103: analyzing the keyword dataset to obtain core keywords and core weights of the core keywords; and Step S104: generating titles based on the core weights of the core keywords.

[0122] According to the technical solution of the embodiment of the present invention, search keywords are extracted from historical search data; search keywords are clustered to obtain a keyword data set; the keyword data set is analyzed to obtain core keywords and core weights of the core keywords; and a technical means of generating titles based on the core weights of the core keywords is used. Titles are generated according to real user operation behaviors, thereby overcoming technical problems such as inaccurate titles, long or redundant texts, low hit conversion rates, and poor user experience, thereby achieving the goal of improving hit conversion rates and user experience; the generated titles accurately include product features that users are concerned about and are consistent with the content of the products.

[0123] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A method for generating a title, characterized in that: include: Extract search keywords from historical search data; Clustering the search keywords to obtain a keyword data set; Analyzing the keyword data set to obtain core keywords and core weights of the core keywords includes: calculating an average value of the keyword data set as the core keyword; assigning initial weights to the core keywords; calculating similarity values ​​between the keyword data sets using a bidirectional neural network; adjusting the initial weights based on the similarity values ​​to obtain adjusted weights, and selecting the largest adjusted weight as the core weight; A title is generated based on the core weight of the core keyword.

2. The method according to claim 1, characterized in that Adjusting the initial weight based on the similarity value to obtain the adjusted weight includes: Adjusting the initial weights using a weight processing formula to obtain adjusted weights, and selecting the largest adjusted weight as the core weight; The weight processing formula is: W=v+α(xv); W is the adjusted weight, v is the initial weight, x is the adjusted independent variable, and α is the similarity value.

3. The method according to claim 1, characterized in that Clustering the search keywords includes: The search keywords are clustered using a K-means algorithm, a K-medoid algorithm, a density-based clustering algorithm, or a Gaussian mixture model.

4. The method according to claim 1, wherein Clustering the search keywords to obtain a keyword data set includes: Randomly select K search keywords as cluster centroids; Dividing the search keywords of the same category as the cluster centroid points into a keyword cluster; Calculate the average value of the keyword cluster as the centroid of the new cluster; Re-dividing the search keywords of the same category as the new cluster centroid points into a keyword cluster; The keyword clusters whose cluster centroids no longer change or whose division times reach a preset value are taken as keyword data sets.

5. The method according to claim 1, wherein Extracting search keywords from historical search data includes: Cluster the historical search data to obtain a historical data set; An average value of the historical data set is calculated as a search keyword.

6. The method according to any one of claims 1 to 5, characterized in that: The method further comprises: Extract real-time keywords based on real-time search data; Clustering the real-time keywords to obtain a real-time data set; Analyzing the real-time data set to obtain a verification keyword and a verification weight of the verification keyword; The title is checked or adjusted based on the check weight of the check keyword.

7. A device for generating a title, characterized in that: include: An extraction module, used to extract search keywords from historical search data; A clustering module, configured to cluster the search keywords to obtain a keyword data set; an analysis module configured to analyze the keyword data set to obtain core keywords and core weights of the core keywords, including: calculating an average value of the keyword data set as the core keyword; assigning an initial weight to the core keyword; calculating similarity values ​​between the keyword data sets using a bidirectional neural network; adjusting the initial weights based on the similarity values ​​to obtain adjusted weights, and selecting the largest adjusted weight as the core weight; A generating module is used to generate a title based on the core weight of the core keyword.

8. The device according to claim 7, characterized in that The analysis module is further configured to: Adjusting the initial weights using a weight processing formula to obtain adjusted weights, and selecting the largest adjusted weight as the core weight; The weight processing formula is: W=v+α(xv); W is the adjusted weight, v is the initial weight, x is the adjusted independent variable, and α is the similarity value.

9. The device according to claim 7, characterized in that The clustering module is also used to: The search keywords are clustered using a K-means algorithm, a K-medoid algorithm, a density-based clustering algorithm, or a Gaussian mixture model.

10. The device according to claim 7, characterized in that The clustering module is also used to: Randomly select K search keywords as cluster centroids; Dividing the search keywords of the same category as the cluster centroid points into a keyword cluster; Calculate the average value of the keyword cluster as the centroid of the new cluster; Re-dividing the search keywords of the same category as the new cluster centroid points into a keyword cluster; The keyword clusters whose cluster centroids no longer change or whose division times reach a preset value are taken as keyword data sets.

11. The device according to claim 7, characterized in that The extraction module is further configured to: Cluster the historical search data to obtain a historical data set; An average value of the historical data set is calculated as a search keyword.

12. The device according to any one of claims 7 to 11, characterized in that: The device further comprises: The verification module is used to extract real-time keywords based on real-time search data; cluster the real-time keywords to obtain a real-time data set; analyze the real-time data set to obtain verification keywords and verification weights of the verification keywords; and verify or adjust the title based on the verification weights of the verification keywords.

13. An electronic device for generating a title, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 6.

14. A computer-readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images