Partner enterprise identification method and device

Abstract

The present invention provides a method and device for identifying cooperative enterprises, which belongs to the field of artificial intelligence technology and can be applied to the field of financial technology or other technical fields. The cooperative enterprise identification method includes: determining a hot search keyword set and a retrieval keyword set according to the matching results of the vocabulary of each hot search summary and the hot search word library; integrating each hot search summary according to the similarity between the keyword sets of each hot search summary, and obtaining the public opinion data corresponding to the integrated hot search summary; retrieving the retrieval keyword set to obtain enterprise data, inputting the public opinion data and enterprise data into a prediction model created based on public opinion training data, enterprise training data and prediction result data, and obtaining enterprise cooperation prediction results; identifying cooperative enterprises according to the enterprise cooperation prediction results. The present invention can grasp the cooperation opportunities in time and effectively improve the cooperation benefits.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method and device for identifying a cooperative enterprise. Background Art

[0002] In the face of fierce market competition, as a product marketing strategy, banks often choose some cooperative institutions to launch co-branded products, such as theme credit cards or theme certificates of deposit, in order to enhance the value of the products by leveraging the influence of both parties. At present, when selecting cooperative enterprises, they often rely on the intuitive judgment of managers on the enterprises, lacking data support and comprehensive and integrated system automatic evaluation solutions, and fail to grasp the opportunities of cooperating with external institutions by leveraging public opinion hotspots in a timely manner. Summary of the invention

[0003] The main purpose of the embodiments of the present invention is to provide a method and device for identifying cooperative enterprises, so as to timely grasp the cooperation opportunities and effectively improve the cooperation benefits.

[0004] In order to achieve the above object, an embodiment of the present invention provides a method for identifying a cooperative enterprise, comprising:

[0005] Determine a hot search keyword set and a search keyword set according to the matching results of the vocabulary of each hot search summary and the hot search word library;

[0006] Integrate the hot search summaries according to the similarity between the keyword sets of the hot search summaries, and obtain the public opinion data corresponding to the integrated hot search summaries;

[0007] Retrieve the search keyword set to obtain enterprise data, input the public opinion data and enterprise data into a prediction model created based on the public opinion training data, the enterprise training data and the prediction result data, and obtain the enterprise cooperation prediction result;

[0008] Identify cooperative enterprises based on enterprise cooperation forecast results.

[0009] The embodiment of the present invention further provides a cooperative enterprise identification device, comprising:

[0010] A set determination module, used to determine a hot search keyword set and a search keyword set according to the matching results of the vocabulary of each hot search summary and the hot search word library;

[0011] The public opinion data module is used to integrate the hot search summaries according to the similarity between the keyword sets of the hot search summaries, and obtain the public opinion data corresponding to the integrated hot search summaries;

[0012] A prediction module, which is used to retrieve a search keyword set to obtain enterprise data, input the public opinion data and enterprise data into a prediction model created based on public opinion training data, enterprise training data and prediction result data, and obtain enterprise cooperation prediction results;

[0013] The identification module is used to identify the cooperating enterprises according to the enterprise cooperation prediction results.

[0014] An embodiment of the present invention further provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the cooperative enterprise identification method when executing the computer program.

[0015] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the cooperative enterprise identification method are implemented.

[0016] The cooperative enterprise identification method and device of the embodiment of the present invention first determine the hot search keyword set and the retrieval keyword set based on the matching results of the vocabulary of each hot search summary and the hot search vocabulary library, and then obtain the corresponding public opinion data after integrating each hot search summary according to the similarity between the keyword sets, and then obtain the enterprise data according to the retrieval keyword set and obtain the enterprise cooperation prediction results based on the public opinion data and the enterprise data to identify the cooperative enterprise, which can timely grasp the cooperation opportunities and effectively improve the cooperation benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 is a flow chart of a method for identifying a cooperative enterprise in an embodiment of the present invention;

[0019] Figure 2 is a flow chart of a method for identifying a cooperative enterprise in another embodiment of the present invention;

[0020] Figure 3 It is a corresponding relationship diagram of hot search titles, news titles and news comments;

[0021] Figure 4 is a flow chart of S101 in an embodiment of the present invention;

[0022] Figure 5 is a flow chart of obtaining public opinion data in an embodiment of the present invention;

[0023] Figure 6 is a structural block diagram of a cooperative enterprise identification device in an embodiment of the present invention;

[0024] Figure 7 It is a structural block diagram of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] Those skilled in the art will appreciate that the embodiments of the present invention may be implemented as a system, device, apparatus, method or computer program product. Therefore, the present disclosure may be specifically implemented in the following forms, namely: complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0027] In view of the fact that the existing technology often relies on the intuitive judgment of managers on enterprises when screening cooperative enterprises, lacks data support and a comprehensive and integrated system automatic evaluation solution, and fails to grasp the opportunity of cooperating with external institutions by taking advantage of public opinion hot spots in a timely manner, the embodiment of the present invention provides a cooperative enterprise identification method and device, which can be based on hot news data, and use artificial intelligence technology to model the public opinion dynamics of news topics, the market, risks, and relationship with banks of news-related institutions in multiple dimensions, so as to evaluate the benefits of co-branded products to banks. The present invention is described in detail below in conjunction with the accompanying drawings.

[0028] Figure 1 4 is a flow chart of a method for identifying a cooperative enterprise in an embodiment of the present invention. Figure 2 FIG. 1 is a flow chart of a method for identifying a cooperative enterprise in another embodiment of the present invention. Figure 1-Figure 2 As shown, the methods for identifying cooperative enterprises include:

[0029] S101: Determine a hot search keyword set and a search keyword set according to the matching results of the vocabulary of each hot search summary and the hot search word library.

[0030] Before executing S101, the system first uses crawler technology to regularly collect news hot search lists released by major news information platforms inside and outside the industry through preset collection channels, collection strategies and collection elements. The collection channel refers to the target news platform that needs to be collected. The collection strategy is to use crawler technology to obtain various parameters preset in advance for news, including collection cycle, collection method (including web pages, public accounts and apps, etc.), collection path (referring to the access path through which the news list is obtained after entering the news platform) and the number of hot search list news collected. Collection elements refer to the content that needs to be collected for each news hot search list, including collection time, collection platform, hot search title, news title, news text, news comments, news likes, news negative reviews and hot search words, etc. Regularly means collecting the above information according to a certain time period. The collection period can be by day, week, month, etc. (weekly is recommended).

[0031] Figure 3 is the corresponding relationship diagram of hot search titles, news titles and news comments. Figure 3 As shown in the figure, the same hot search title is often associated with one or more news titles, and the content of the news is highly related to the hot search title; and the same hot search news often has multiple news comments. The hot search title, news title and news comment are in a tree-like relationship. Figure 3 Where h and j are positive integers.

[0032] For example, a hot search title on the news hot search list is "** officially opened". After clicking on this title, you can get multiple news related to the hot search title, such as "** officially opened, tickets are hotly snapped up", "A family of three can easily spend 10,000 yuan on a trip to **...", "** Avenue, which does not require tickets, is very popular on weekends...", etc. Through each news title, you can view the news text and news comments of the title.

[0033] In the next step, the news titles acquired in a time period can be summarized according to their association with the hot search titles. For example, all news titles associated with the same hot search title can be summarized into the same text (i.e., different news titles are connected end to end) as the news titles used in the subsequent steps, i.e., the news titles mentioned in the subsequent steps refer to the summarized news titles.

[0034] Similarly, the number of likes and negative reviews for all news items associated with the same hot search title are summarized (i.e., the number of likes for different news titles are added up, and the number of negative reviews for different news titles are added up). The number of likes and negative reviews for news items mentioned in subsequent steps refer to the summarized values.

[0035] It should be noted that since each news comment may express different emotions, it is necessary to analyze and identify emotions separately, so the present invention does not aggregate news comments. In addition, the subsequent steps need to use hot search titles, news titles, news likes, news dislikes and news comments. The rest of the information obtained by the crawler (such as collection time, collection platform, news text, etc.) can be saved in the system as news details for manual query.

[0036] In one embodiment, before executing S101, the method further includes:

[0037] Generate original hot search data according to the hot search titles and corresponding news titles; input the original hot search data into a summary recognition model created based on historical summary data to obtain a hot search summary.

[0038] In the specific implementation, each hot search title and its associated news title are summarized into the same text (that is, the hot search title and the news title are connected end to end), and defined as the original hot search data. The original hot search data is first segmented through the open source word segmentation library, and then the summary recognition model created based on the historical summary data is used to determine whether it belongs to business news. Whether it belongs to business news is a binary classification problem. The commonly used machine learning supervised learning algorithms in the industry can be used to train classification models, such as the Naive Bayes algorithm, the SVM algorithm, and the LSTM algorithm. Through model training, a summary recognition model for determining whether it is business news can be obtained. After obtaining new original hot search data, the classification model can be used to identify whether each original hot search data is business news. For news that does not belong to business news, it will not be processed in the subsequent steps. The system only retains the original hot search data corresponding to the business news as the hot search summary for subsequent processing.

[0039] Figure 4 is a flow chart of S101 in an embodiment of the present invention. Figure 4 As shown, S101 includes:

[0040] S201: Determine adjustment parameters of each word according to the matching results between the words in each hot search summary and the hot search word library.

[0041] Among them, hot search words refer to popular search words. The system removes the hot search words and retains them in the local hot search word library for processing in subsequent steps. In specific implementation, each word in the hot search summary can be matched with the hot search word library. If it matches the words in the hot search word library, it is considered to hit the hot search word library and use the adjustment parameter α. Otherwise, it does not hit the hot search word library.

[0042] S202: Determine the frequency data of each word according to the adjustment parameter, word frequency and frequency index of each word.

[0043] The present invention calculates the frequency data of each word in the hot search summary through the TF-IDF word frequency analysis method. The greater the contribution of a word to reflecting the unique theme of the article, the higher the frequency data. In specific implementation, the frequency data can be determined by the following formula:

[0044]

[0045] Among them, TFIDF new is the frequency data of the vocabulary, TF is the word frequency of the vocabulary, and IDF is the inverse document frequency index of the vocabulary; α and α -1 are both adjustment parameters. When the vocabulary matches a word in the hot search term library, the adjustment parameter is α. When the vocabulary does not match a word in the hot search term library, the adjustment parameter is α -1 , and usually α≥1.

[0046] S203: Determine the hot search keyword set and the retrieval keyword set according to the frequency data of each vocabulary.

[0047] In specific implementation, the N vocabulary with the highest TFIDF new in each hot search summary can be extracted as the keyword set of this hot search summary (N can be taken between 20 and 50 in actual use); the Np vocabulary with the highest TFIDF new in each hot search summary are extracted as the retrieval keyword set (Np < N, and the value of Np is relatively small, such as 3).

[0048] S102: Integrate each hot search summary according to the similarity between the keyword sets of each hot search summary, and obtain the public opinion data corresponding to the integrated hot search summary.

[0049] Among them, the integration of the hot search summary includes the integration of the hot search summaries within the same period, as well as the integration of the hot search summaries in the current period and the historical period.

[0050] In one embodiment, integrating each hot search summary according to the similarity between the keyword sets of each hot search summary includes:

[0051] Converting the keyword sets of each hot search summary into word vectors; when the similarity between the word vectors meets the preset similarity condition, integrating the corresponding hot search summaries.

[0052] In specific implementation, the keywords in the keyword set can be converted into K-dimensional word vectors through an open source pre-trained word vector model (K can be a power of 2 such as 128, 256, or other integers in actual use), and the word vector summary value of each abstract can be summarized. Calculate the cosine similarity of the word vector summary values ​​of the two hot search summaries. The larger the cosine similarity value, the more similar the two hot search summaries are. A threshold M can be set (M can be a value between 0.6-0.9 in actual use) to determine that the hot search summaries greater than M are content of the same topic and integrate them. Conversely, hot search summaries less than or equal to M are determined to be content of different topics and are not integrated. N, K, M, and Np are all adjustable parameters. In actual applications, the initial values ​​can be set according to manual experience, and then adjusted according to the actual model training effect.

[0053] For example, "Today ** officially opens!", "**: Tourists go straight to "##"" and "%% cancels 220 million shares and stands with investors." The three titles are hot search titles obtained from hot search lists on different platforms, but the first two are hot search titles related to "**" and are similar hot searches, so they can be integrated. The third one is obviously different from the first two and needs to be treated as a separate hot search title, and is not integrated with the first two.

[0054] Figure 5 FIG. 1 is a flow chart of obtaining public opinion data in an embodiment of the present invention. Figure 5 As shown, the public opinion data corresponding to the integrated hot search summary includes:

[0055] S301: Determine public opinion index data according to the sentiment index data of the comments corresponding to the integrated hot search summary.

[0056] Among them, public opinion indicator data include public opinion parameter data and public opinion fluctuation data.

[0057] In one embodiment, S301 includes:

[0058] Determine the public opinion parameter data based on the sentiment index data of the comments corresponding to the integrated hot search summary;

[0059] Determine public opinion fluctuation data based on sentiment indicator data and public opinion parameter data.

[0060] In the specific implementation, it is necessary to first analyze the sentiment index data of each news comment before integration. Sentiment index data includes five categories: "0-strong support, 1-support, 2-neutral, 3-complaint, 4-strong complain". Therefore, the problem of sentiment index data is actually a multi-classification problem. The sentiment analysis supervised learning algorithms commonly used in the industry can be used to train classification models, such as the text convolutional neural network model (TextCNN), BERT algorithm and naive Bayes algorithm. Each public opinion evaluation can be regarded as an entity. Through the sentiment analysis supervised learning algorithm, each comment is sentimentally classified, and the sentiment index of each comment is obtained, so that the result of each public opinion evaluation can be identified.

[0061] After obtaining the result of each public opinion evaluation, the hot search titles to which the news title belongs can be summarized and counted, thereby calculating the number of hot search titles evaluated as 0-strong support, 1-support, 2-neutral, 3-complaint and 4-strong complain, respectively recorded as Ni (i=0-4). The public opinion parameter data is the average value of the sentiment index data, and the public opinion fluctuation data is the variance of the sentiment index data.

[0062] When integrating the hot search summaries of the current cycle and the historical cycle, similarity analysis is performed on the hot search summaries of the current cycle and the hot search summaries of the historical cycle. If no result with high similarity to the hot search summaries of the current cycle is found in the hot search summaries of the historical cycle, the hot search summaries of the current cycle are used as an instance. On the contrary, if it is found that the hot search summary of the current cycle is highly similar to the hot search summary of a certain historical cycle (if there are multiple historical hot search summaries with high similarity, the highest historical hot search summary is taken), the sentiment index data of the hot search summary of the current cycle is used as the data of a certain statistical time period of the historical hot search summary.

[0063] Table 1

[0064]

[0065] Table 1 is a time series data table of public opinion indicators in the first embodiment. As shown in Table 1, the historical hot search news is "** opens today", and the public opinion parameter data and public opinion fluctuation data of week T are 4.2 and 0.02 respectively. The hot search news of week T+1 is "** holiday guide", and its public opinion parameter data and public opinion fluctuation data are 4.0 and 0.01 respectively. If the similarity analysis finds that the similarity between the two is high, a time series data will be generated in the end.

[0066] Table 2

[0067]

[0068] Table 2 is a time series data table of public opinion indicators in the second embodiment. As shown in Table 2, the historical hot search news is "** opens today", and the public opinion parameter data and public opinion fluctuation data of week T are 4.2 and 0.02 respectively. The hot search news of week T+1 is "×× event accepts reservations", and its public opinion parameter data and public opinion fluctuation data are 4.4 and 0.01 respectively. If the similarity analysis finds that the similarity between the two is low, and the similarity between other hot search news and "** opens today" is also low, then "**" and "××" will generate two time series data, and it is necessary to fill 0 for "**" of week T+1; and fill 0 for "××" of week T.

[0069] If it is found that the current hot search summary has a high similarity with a historical hot search summary, in addition to summarizing, it is necessary to further update the word vector of the historical hot search summary. The update formula is:

[0070] The updated word vector of the historical hot search summary = (the word vector of the historical hot search summary + the word vector of the current hot search summary) / 2.

[0071] S302: Determine the news like rate according to the like data of the news corresponding to the integrated hot search summary.

[0072] Among them, the likes data includes the number of news likes and the number of news negative reviews. The number of news likes is the sum of the number of news likes associated with the hot search titles to be integrated on different platforms. Similarly, the number of news negative reviews is the sum of the number of news negative reviews on different platforms in the same way. News like rate = news likes / (news likes + news negative reviews).

[0073] S303: Obtain the public opinion platform data corresponding to the integrated hot search summary.

[0074] Among them, the public opinion platform data is the number of hot search summaries entering the hot search lists of different platforms.

[0075] S103: Retrieve the search keyword set to obtain enterprise data, input the public opinion data and enterprise data into a prediction model created based on the public opinion training data, enterprise training data and prediction effect data, and obtain enterprise cooperation prediction results.

[0076] Among them, the public opinion training data is time series data, and the enterprise training data and prediction effect data are non-time series data (i.e. historical point data). Therefore, the time series data needs to be converted into non-time series data. Usually, new statistical indicators can be established according to different time periods, such as the public opinion training data of the past 1 month, the public opinion training data of the past 3 months, and the highest value of the public opinion indicator data of the past 3 months.

[0077] Table 3

[0078]

[0079] Table 4

[0080]

[0081] Table 5

[0082]

[0083] Table 3 is a time series table of public opinion training data, Table 4 is a time series table of public opinion training specific data, and Table 5 is a non-time series table of public opinion training specific data. As shown in Tables 3-5, after converting the time series data into non-time series data, the model needs to predict the enterprise cooperation prediction results through public opinion training data, enterprise training data and prediction result data. This problem belongs to the regression model of machine learning. The commonly used algorithms in the industry such as GBRFR (gradient boosted random forest regression) and ETR (EXTRA TREE regression) can be used to train regression models.

[0084] For example, for a certain hot search news, the model estimates the number of credit cards issued in the next three months based on the public opinion statistics of the hot search news in the past six months and the corporate data at the current time point. For example, when the system obtains hot search news related to "**", it can estimate the number of credit cards issued in the next three months for issuing ** credit cards.

[0085] This application can also apply the Wide and Deep neural network model to conduct modeling based on both time series data and non-time series data. In specific implementation, the time series data can be used as the Deep feature, and the Wide feature can be associated with the non-time series data through a shallow fully connected network. The enterprise cooperation prediction result still uses a single indicator value.

[0086] In one embodiment, searching a search keyword set to obtain enterprise data includes:

[0087] The company name is obtained by searching the search keyword set in the cooperation system; and the company data is obtained according to the company name and the company cooperation prediction result definition.

[0088] In specific implementation, the names of enterprises can be searched in enterprise information retrieval systems (such as Tianyancha and enterprise registration websites) based on Np words. new The top three words are "**", "price" and "queue". By searching for "**", you can find specific company names. In order to avoid searching for multiple companies with different keywords, the search terms can be manually calibrated.

[0089] Enterprise data includes basic data on the comprehensive influence of enterprises and service relationship data of enterprises. Enterprise scale, enterprise credit data and enterprise operating status are all related to whether banks will choose to cooperate with them. In order to avoid market risks and reputation risks brought by joint cooperation products to banks, they often choose to cooperate with enterprises with good credit and good operating conditions. The basic data on the comprehensive influence of enterprises, such as enterprise scale, enterprise credit status and number of legal proceedings, can be obtained in the enterprise information retrieval system based on the enterprise name through crawler technology.

[0090] Table 6

[0091]

[0092] Table 6 is a service relationship data table. As shown in Table 6, the service relationship elements of an enterprise include data such as assets, intermediate income contribution, and asset sedimentation of corporate legal persons. The data can be preprocessed to be suitable for subsequent machine learning models, such as grouping continuous variables to convert them into discrete variables, and normalizing discrete variables.

[0093] The definition of the enterprise cooperation prediction results can be determined based on the core product operation indicators of different joint cooperation products, and appropriate enterprise data can be selected for training based on the definition. For example, if the joint cooperation product is a credit card product, for banks, the core product operation indicators are usually the number of credit cards issued or the total consumption of credit card customers. Therefore, the enterprise cooperation prediction results can be defined as the number of credit cards issued or the total consumption of credit card customers; if the core product operation indicator is a large-denomination certificate of deposit product, the enterprise cooperation prediction results can be defined as the total purchase amount of certificate of deposit customers. Specifically, it is necessary to determine the definition of the enterprise cooperation prediction results in advance and determine the statistical period of the prediction. For example, the enterprise cooperation prediction results are defined as the number of credit cards issued in the three months after the joint venture.

[0094] S104: Identify cooperative enterprises based on the enterprise cooperation prediction results.

[0095] In specific implementation, by achieving a set threshold Nk (Nk is the preset minimum marketing target value of the bank), the hot search news and company names with a credit card issuance volume greater than Nk in the next three months can be pushed to bank managers as cooperative companies, providing a decision-making basis for bank managers to carry out joint product cooperation with companies.

[0096] Figure 1 The execution subject of the cooperative enterprise identification method shown can be a computer. Figure 1As can be seen from the process shown, the method for identifying cooperative enterprises in an embodiment of the present invention first determines a hot search keyword set and a retrieval keyword set based on the matching results of the vocabulary of each hot search summary and the hot search vocabulary library, and then obtains the corresponding public opinion data after integrating each hot search summary according to the similarity between the keyword sets, and then obtains the enterprise data based on the retrieval keyword set and obtains the enterprise cooperation prediction results based on the public opinion data and the enterprise data to identify the cooperative enterprises, which can timely grasp the cooperation opportunities and effectively improve the cooperation benefits.

[0097] The specific process of the embodiment of the present invention is as follows:

[0098] 1. Generate original hot search data based on hot search titles and corresponding news titles.

[0099] 2. Input the original hot search data into the summary recognition model created based on the historical summary data to obtain the hot search summary.

[0100] 3. Determine the adjustment parameters of each word according to the matching results of the words in each hot search summary and the hot search word library.

[0101] 4. Determine the frequency data of each word based on its adjustment parameters, word frequency and frequency index.

[0102] 5. Determine the hot search keyword set and the search keyword set based on the frequency data of each word.

[0103] 6. Convert the keyword set of each hot search summary into a word vector. When the similarity between the word vectors meets the preset similarity condition, integrate the corresponding hot search summary.

[0104] 7. Determine public opinion parameter data based on the sentiment index data of the comments corresponding to the integrated hot search summary, and determine public opinion fluctuation data based on the sentiment index data and public opinion parameter data.

[0105] 8. Determine the news like rate based on the like data of the news corresponding to the integrated hot search summary, and obtain the public opinion platform data corresponding to the integrated hot search summary.

[0106] 9. Search the keyword set in the cooperation system to obtain the company name, and obtain the company data based on the company name and the company cooperation prediction result definition.

[0107] 10. Input the public opinion data and enterprise data into the prediction model created based on the public opinion training data, enterprise training data and prediction result data to obtain the enterprise cooperation prediction results, and identify the cooperating enterprises based on the enterprise cooperation prediction results.

[0108] In summary, the cooperative enterprise identification method of the embodiment of the present invention is based on hot news data. It uses artificial intelligence algorithm technology to comprehensively model the public opinion dynamics of news topics, the market, risks, and relationship with banks of news-related institutions, etc., so as to evaluate the benefits brought to the bank by co-branded products, help managers better grasp the opportunities to use public opinion hotspots to develop cooperation with external institutions, and bring better cooperation benefits.

[0109] Based on the same inventive concept, an embodiment of the present invention further provides a cooperative enterprise identification device. Since the principle of solving the problem by the device is similar to that of the cooperative enterprise identification method, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0110] Figure 6 : is a structural block diagram of a cooperative enterprise identification device in an embodiment of the present invention. Figure 6 As shown, the cooperative enterprise identification device includes:

[0111] A set determination module, used to determine a hot search keyword set and a search keyword set according to the matching results of the words of each hot search summary and the hot search word library;

[0112] The public opinion data module is used to integrate the hot search summaries according to the similarity between the keyword sets of the hot search summaries, and obtain the public opinion data corresponding to the integrated hot search summaries;

[0113] A prediction module, which is used to retrieve a search keyword set to obtain enterprise data, input the public opinion data and enterprise data into a prediction model created based on public opinion training data, enterprise training data and prediction result data, and obtain enterprise cooperation prediction results;

[0114] The identification module is used to identify the cooperating enterprises according to the enterprise cooperation prediction results.

[0115] In summary, the cooperative enterprise identification device of the embodiment of the present invention first determines the hot search keyword set and the retrieval keyword set according to the matching results of the vocabulary of each hot search summary and the hot search vocabulary library, and then obtains the corresponding public opinion data after integrating each hot search summary according to the similarity between the keyword sets, and then obtains the enterprise data according to the retrieval keyword set and obtains the enterprise cooperation prediction results according to the public opinion data and the enterprise data to identify the cooperative enterprises, which can timely grasp the cooperation opportunities and effectively improve the cooperation benefits.

[0116] The embodiment of the present invention also provides a specific implementation of a computer device capable of implementing all the steps in the cooperative enterprise identification method in the above embodiment. Figure 7 is a block diagram of a computer device in an embodiment of the present invention, see Figure 7 , the computer equipment specifically includes the following contents:

[0117] Processor (processor) 701 and memory (memory) 702.

[0118] The processor 701 is used to call the computer program in the memory 702. When the processor executes the computer program, all steps in the cooperative enterprise identification method in the above embodiment are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0119] Determine a hot search keyword set and a search keyword set according to the matching results of the vocabulary of each hot search summary and the hot search word library;

[0120] Integrate the hot search summaries according to the similarity between the keyword sets of the hot search summaries, and obtain the public opinion data corresponding to the integrated hot search summaries;

[0121] Retrieve the search keyword set to obtain enterprise data, input the public opinion data and enterprise data into a prediction model created based on the public opinion training data, the enterprise training data and the prediction result data, and obtain the enterprise cooperation prediction result;

[0122] Identify cooperative enterprises based on enterprise cooperation forecast results.

[0123] In summary, the computer device of the embodiment of the present invention first determines the hot search keyword set and the retrieval keyword set according to the matching results of the vocabulary of each hot search summary and the hot search vocabulary library, and then obtains the corresponding public opinion data after integrating each hot search summary according to the similarity between the keyword sets, and then obtains the enterprise data according to the retrieval keyword set and obtains the enterprise cooperation prediction results according to the public opinion data and the enterprise data to identify the cooperating enterprises, so as to seize the cooperation opportunities in time and effectively improve the cooperation benefits.

[0124] The embodiment of the present invention also provides a computer-readable storage medium capable of implementing all the steps in the cooperative enterprise identification method in the above embodiment. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, all the steps in the cooperative enterprise identification method in the above embodiment are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0125] Determine a hot search keyword set and a search keyword set according to the matching results of the vocabulary of each hot search summary and the hot search word library;

[0126] Integrate the hot search summaries according to the similarity between the keyword sets of the hot search summaries, and obtain the public opinion data corresponding to the integrated hot search summaries;

[0127] Retrieve the search keyword set to obtain enterprise data, input the public opinion data and enterprise data into a prediction model created based on the public opinion training data, the enterprise training data and the prediction result data, and obtain the enterprise cooperation prediction result;

[0128] Identify cooperative enterprises based on enterprise cooperation forecast results.

[0129] In summary, the computer-readable storage medium of the embodiment of the present invention first determines the hot search keyword set and the retrieval keyword set according to the matching results of the vocabulary of each hot search summary and the hot search vocabulary, and then obtains the corresponding public opinion data after integrating each hot search summary according to the similarity between the keyword sets, and then obtains the enterprise data according to the retrieval keyword set and obtains the enterprise cooperation prediction results according to the public opinion data and the enterprise data to identify the cooperative enterprises, so as to seize the cooperation opportunities in time and effectively improve the cooperation benefits.

[0130] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0131] Those skilled in the art may also understand that the various illustrative logical blocks, units, and steps listed in the embodiments of the present invention may be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly demonstrate the interchangeability of hardware and software, the various illustrative components, units, and steps described above have generally described their functions. Whether such functions are implemented by hardware or software depends on the specific application and the design requirements of the entire system. Those skilled in the art may use various methods to implement the described functions for each specific application, but such implementation should not be understood as exceeding the scope of protection of the embodiments of the present invention.

[0132] The various illustrative logic blocks, or units, or devices described in the embodiments of the present invention can be implemented or operated by a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array or other programmable logic device, a discrete gate or transistor logic, a discrete hardware component, or any combination of the above. The general-purpose processor can be a microprocessor, and optionally, the general-purpose processor can also be any conventional processor, controller, microcontroller or state machine. The processor can also be implemented by a combination of computing devices, such as a digital signal processor and a microprocessor, a plurality of microprocessors, one or more microprocessors combined with a digital signal processor core, or any other similar configuration.

[0133] The steps of the method or algorithm described in the embodiments of the present invention can be directly embedded in hardware, a software module executed by a processor, or a combination of the two. The software module can be stored in a RAM memory, a flash memory, a ROM memory, an EPROM memory, an EEPROM memory, a register, a hard disk, a removable disk, a CD-ROM, or other storage media of any form in the art. Exemplarily, the storage medium can be connected to the processor so that the processor can read information from the storage medium and can write information to the storage medium. Optionally, the storage medium can also be integrated into the processor. The processor and the storage medium can be arranged in an ASIC, and the ASIC can be arranged in a user terminal. Optionally, the processor and the storage medium can also be arranged in different components in the user terminal.

[0134] In one or more exemplary designs, the above functions described in the embodiments of the present invention can be implemented in hardware, software, firmware or any combination of the three. If implemented in software, these functions can be stored on a computer-readable medium, or transmitted in the form of one or more instructions or codes on a computer-readable medium. Computer-readable media include computer storage media and communication media that facilitate the transfer of computer programs from one place to another. The storage medium can be any available medium that can be accessed by any general or special computer. For example, such computer-readable media can include but are not limited to RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage devices, or any other medium that can be used to carry or store program codes in the form of instructions or data structures and other forms that can be read by general or special computers, or general or special processors. In addition, any connection can be appropriately defined as a computer-readable medium, for example, if the software is transmitted from a website site, server or other remote resource through a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL) or wirelessly, such as infrared, wireless and microwave, it is also included in the defined computer-readable medium. The disk and disc include compact disk, laser disk, optical disk, DVD, floppy disk and blue-ray disk. Disks usually copy data magnetically, while discs usually copy data optically with lasers. The above combination can also be included in computer readable media.

Claims

1. A method for identifying a cooperative enterprise, characterized in that: include: Generate original hot search data based on hot search titles and corresponding news titles; Inputting the original hot search data into a summary recognition model created based on the historical summary data to obtain a hot search summary; Determine a hot search keyword set and a search keyword set according to the matching results of the vocabulary of each hot search summary and the hot search word library; Integrate the hot search summaries according to the similarity between the keyword sets of the hot search summaries, and obtain the public opinion data corresponding to the integrated hot search summaries; Retrieving the search keyword set to obtain enterprise data, inputting the public opinion data and the enterprise data into a prediction model created based on public opinion training data, enterprise training data and prediction result data, and obtaining enterprise cooperation prediction results, Wherein, searching the search keyword set to obtain enterprise data includes: searching the search keyword set in the cooperation system to obtain enterprise names, and obtaining enterprise data according to the enterprise names and enterprise cooperation prediction result definitions; The cooperating enterprises are identified according to the enterprise cooperation prediction results.

2. The cooperative enterprise identification method according to claim 1, characterized in that: According to the matching results of the words of each hot search summary and the hot search word library, the hot search keyword set and the search keyword set are determined to include: Determine the adjustment parameters of each word according to the matching results of the words in each hot search summary and the hot search word library; Determining frequency data for each word based on the adjustment parameters, word frequency and frequency index of each word; Determine the hot search keyword set and the search keyword set based on the frequency data of each word.

3. The cooperative enterprise identification method according to claim 1, characterized in that: Integrating each hot search summary based on the similarity between the keyword sets of each hot search summary includes: Convert the keyword set of each hot search summary into a word vector; When the similarity between word vectors meets the preset similarity condition, the corresponding hot search summaries are integrated.

4. The cooperative enterprise identification method according to claim 1, characterized in that: The public opinion data corresponding to the integrated hot search summary includes: Determine the public opinion index data based on the sentiment index data of the comments corresponding to the integrated hot search summary; Determine the news like rate based on the like data of the news corresponding to the integrated hot search summary; Get the public opinion platform data corresponding to the integrated hot search summary.

5. The cooperative enterprise identification method according to claim 4, characterized in that: The public opinion index data includes public opinion parameter data and public opinion fluctuation data; According to the sentiment index data of the comments corresponding to the integrated hot search summary, the public opinion index data includes: Determine the public opinion parameter data based on the sentiment index data of the comments corresponding to the integrated hot search summary; The public opinion fluctuation data is determined according to the sentiment index data and the public opinion parameter data.

6. A cooperative enterprise identification device, characterized in that: include: A generation module, used to generate original hot search data according to hot search titles and corresponding news titles; An input module, used for inputting the original hot search data into a summary recognition model created based on the historical summary data to obtain a hot search summary; A set determination module, used to determine a hot search keyword set and a search keyword set according to the matching results of the words of each hot search summary and the hot search word library; The public opinion data module is used to integrate the hot search summaries according to the similarity between the keyword sets of the hot search summaries, and obtain the public opinion data corresponding to the integrated hot search summaries; A prediction module is used to retrieve the search keyword set to obtain enterprise data, input the public opinion data and the enterprise data into a prediction model created based on public opinion training data, enterprise training data and prediction result data, and obtain enterprise cooperation prediction results. The prediction module includes: an acquisition submodule, which is used to retrieve the search keyword set in the cooperation system to obtain the enterprise name; a definition submodule, which is used to obtain enterprise data according to the enterprise name and the enterprise cooperation prediction result definition; An identification module is used to identify cooperative enterprises based on the enterprise cooperation prediction results.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the cooperative enterprise identification method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the cooperative enterprise identification method according to any one of claims 1 to 5 are implemented.

Images