Model-based strategic emerging industry product recommendation method and device, computer equipment, storage medium and computer program product

Abstract

The invention relates to a model-based strategic emerging industry product recommendation method and device, computer equipment, a storage medium and a computer program product. The method comprises the steps of extracting a current keyword in a current index in a current index set of each dimension of a to-be-analyzed product, and determining a current weight corresponding to the current index according to the current keyword; inputting the index data of the current index into the trained evaluation value prediction model to obtain a prediction evaluation value corresponding to the index data, and performing fusion processing on the prediction evaluation value according to the current weight to obtain a fusion evaluation value corresponding to the current index; determining a current product classification result of the to-be-analyzed product according to the fusion evaluation value; and screening out the to-be-analyzed product of which the current product classification result meets the product demand information of the to-be-analyzed object from the to-be-analyzed products, taking the to-be-analyzed product as a target product corresponding to the to-be-analyzed object, and recommending the target product to the to-be-analyzed object. By adopting the method, the accuracy of product recommendation can be improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a model-based strategic emerging industry product recommendation method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Art

[0002] With the emergence of more and more new products (such as products from strategic emerging industries), it is crucial to accurately recommend products.

[0003] In traditional technology, manual judgment is generally used in the process of product recommendation. However, this manual judgment method has subjective factors and is prone to errors, resulting in low accuracy of product recommendations. Summary of the Invention

[0004] Based on this, it is necessary to provide a model-based strategic emerging industry product recommendation method, device, computer equipment, computer-readable storage medium and computer program product that can improve the accuracy of product recommendations in response to the above technical problems.

[0005] First, this application provides a model-based method for recommending products in strategic emerging industries, including:

[0006] Determining a current indicator set of multiple dimensions of a product to be analyzed; the product to be analyzed is used to represent a product output by a strategic emerging industry; the current indicator set of each dimension includes multiple current indicators;

[0007] Extracting current keywords from the current indicators in the current indicator set of each dimension, and determining the current importance corresponding to the current indicator based on the current keywords;

[0008] Determining a current weight corresponding to the current indicator according to the current importance;

[0009] Obtaining indicator data of the current indicator, inputting the indicator data into a trained evaluation value prediction model to obtain a predicted evaluation value corresponding to the indicator data, and fusing the predicted evaluation values according to the current weight to obtain a fused evaluation value corresponding to the current indicator;

[0010] Determining a current product classification result of the product to be analyzed based on the fusion evaluation value;

[0011] Obtain product demand information of the object to be analyzed, filter out the products to be analyzed whose current product classification results meet the product demand information from the products to be analyzed, use them as target products corresponding to the object to be analyzed, and recommend the target products to the object to be analyzed.

[0012] In one embodiment, determining the current importance corresponding to the current indicator based on the current keyword includes:

[0013] Obtain similar keywords corresponding to the current keyword;

[0014] Determining a first importance of the current keyword based on the current keyword, the similar keywords, and the long text associated with the product to be analyzed; determining a second importance of the current keyword based on the current keyword, the similar keywords, and a trained sentiment prediction model; and determining a third importance of the current keyword based on a preset correspondence between the current keyword, the similar keywords, and pre-established keywords and importances;

[0015] The first importance, the second importance, and the third importance are fused to obtain the current importance.

[0016] In one embodiment, determining the first importance of the current keyword based on the current keyword, the similar keywords, and the long text associated with the product to be analyzed includes:

[0017] Determining a first occurrence frequency of the current keyword in the long text and a second occurrence frequency of the similar keyword in the long text;

[0018] performing a fusion process on the first occurrence frequency and the second occurrence frequency to obtain a target occurrence frequency corresponding to the current keyword;

[0019] The first importance is determined according to the target occurrence frequency.

[0020] In one embodiment, determining the second importance of the current keyword based on the current keyword, the similar keywords, and the trained sentiment prediction model includes:

[0021] Inputting the current keyword and the similar keyword into the trained emotion prediction model respectively to obtain a first emotion prediction result of the current keyword and a second emotion prediction result of the similar keyword;

[0022] Querying the correspondence between the emotion prediction results and the importance, and obtaining the fourth importance corresponding to the first emotion prediction result and the fifth importance corresponding to the second emotion prediction result;

[0023] Determining a first weight corresponding to the first emotion prediction result based on a first prediction probability corresponding to the first emotion prediction result output by the trained emotion prediction model, and determining a second weight corresponding to the second emotion prediction result based on a second prediction probability corresponding to the second emotion prediction result output by the trained emotion prediction model;

[0024] The fourth importance and the fifth importance are summed according to the first weight and the second weight to obtain the second importance.

[0025] In one embodiment, determining the third importance of the current keyword based on the preset correspondence between the current keyword, the similar keywords, and the pre-established keywords and importances includes:

[0026] Querying the preset corresponding relationship to obtain the sixth importance corresponding to the current keyword and the seventh importance corresponding to the similar keyword;

[0027] Determining a preset weight corresponding to the similar keyword based on the similarity between the current keyword and the similar keyword;

[0028] The sixth importance and the seventh importance are summed according to the basic weight of the current keyword and the preset weight to obtain the third importance.

[0029] In one embodiment, the current indicator sets of the multiple dimensions include a first indicator set, a second indicator set, and a third indicator set;

[0030] Determining the current product classification result of the product to be analyzed based on the fusion evaluation value includes:

[0031] Determining, based on the fused evaluation value, a first evaluation value of the product to be analyzed under the first indicator set, a second evaluation value of the product to be analyzed under the second indicator set, and a third evaluation value of the product to be analyzed under the third indicator set;

[0032] querying, based on the first evaluation value, the second evaluation value, and the third evaluation value, a correspondence between the evaluation values and the product classification results, to obtain a first product classification result of the product to be analyzed under the first indicator set, a second product classification result of the product to be analyzed under the second indicator set, and a third product classification result of the product to be analyzed under the third indicator set;

[0033] Product categories corresponding to the first product classification result, the second product classification result, and the third product classification result are obtained as the current product classification result.

[0034] In a second aspect, the present application also provides a model-based strategic emerging industry product recommendation device, comprising:

[0035] A set determination module is used to determine a current indicator set of multiple dimensions of a product to be analyzed; the product to be analyzed is used to represent the output of a strategic emerging industry; the current indicator set of each dimension includes multiple current indicators;

[0036] An importance determination module is used to extract current keywords from the current indicators in the current indicator set of each dimension, and determine the current importance corresponding to the current indicator based on the current keywords;

[0037] A weight determination module, configured to determine a current weight corresponding to the current indicator according to the current importance;

[0038] An evaluation value determination module is used to obtain the indicator data of the current indicator, input the indicator data into the trained evaluation value prediction model to obtain the predicted evaluation value corresponding to the indicator data, and fuse the predicted evaluation values according to the current weight to obtain a fused evaluation value corresponding to the current indicator;

[0039] A result determination module, configured to determine a current product classification result of the product to be analyzed based on the fusion evaluation value;

[0040] The product recommendation module is used to obtain product demand information of the object to be analyzed, filter out the products to be analyzed whose current product classification results meet the product demand information from each of the products to be analyzed, use them as target products corresponding to the object to be analyzed, and recommend the target products to the object to be analyzed.

[0041] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0042] Determining a current indicator set of multiple dimensions of a product to be analyzed; the product to be analyzed is used to represent a product output by a strategic emerging industry; the current indicator set of each dimension includes multiple current indicators;

[0043] Extracting current keywords from the current indicators in the current indicator set of each dimension, and determining the current importance corresponding to the current indicator based on the current keywords;

[0044] Determining a current weight corresponding to the current indicator according to the current importance;

[0045] Obtaining indicator data of the current indicator, inputting the indicator data into a trained evaluation value prediction model to obtain a predicted evaluation value corresponding to the indicator data, and fusing the predicted evaluation values according to the current weight to obtain a fused evaluation value corresponding to the current indicator;

[0046] Determining a current product classification result of the product to be analyzed based on the fusion evaluation value;

[0047] Obtain product demand information of the object to be analyzed, filter out the products to be analyzed whose current product classification results meet the product demand information from the products to be analyzed, use them as target products corresponding to the object to be analyzed, and recommend the target products to the object to be analyzed.

[0048] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0049] Determining a current indicator set of multiple dimensions of a product to be analyzed; the product to be analyzed is used to represent a product output by a strategic emerging industry; the current indicator set of each dimension includes multiple current indicators;

[0050] Extracting current keywords from the current indicators in the current indicator set of each dimension, and determining the current importance corresponding to the current indicator based on the current keywords;

[0051] Determining a current weight corresponding to the current indicator according to the current importance;

[0052] Obtaining indicator data of the current indicator, inputting the indicator data into a trained evaluation value prediction model to obtain a predicted evaluation value corresponding to the indicator data, and fusing the predicted evaluation values according to the current weight to obtain a fused evaluation value corresponding to the current indicator;

[0053] Determining a current product classification result of the product to be analyzed based on the fusion evaluation value;

[0054] Obtain product demand information of the object to be analyzed, filter out the products to be analyzed whose current product classification results meet the product demand information from the products to be analyzed, use them as target products corresponding to the object to be analyzed, and recommend the target products to the object to be analyzed.

[0055] In a fifth aspect, the present application further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the following steps:

[0056] Determining a current indicator set of multiple dimensions of a product to be analyzed; the product to be analyzed is used to represent a product output by a strategic emerging industry; the current indicator set of each dimension includes multiple current indicators;

[0057] Extracting current keywords from the current indicators in the current indicator set of each dimension, and determining the current importance corresponding to the current indicator based on the current keywords;

[0058] Determining a current weight corresponding to the current indicator according to the current importance;

[0059] Obtaining indicator data of the current indicator, inputting the indicator data into a trained evaluation value prediction model to obtain a predicted evaluation value corresponding to the indicator data, and fusing the predicted evaluation values according to the current weight to obtain a fused evaluation value corresponding to the current indicator;

[0060] Determining a current product classification result of the product to be analyzed based on the fusion evaluation value;

[0061] Obtain product demand information of the object to be analyzed, filter out the products to be analyzed whose current product classification results meet the product demand information from the products to be analyzed, use them as target products corresponding to the object to be analyzed, and recommend the target products to the object to be analyzed.

[0062] The above-mentioned model-based strategic emerging industry product recommendation method, device, computer equipment, storage medium and computer program product first take the products produced by strategic emerging industries as the products to be analyzed, and determine the current indicator sets of multiple dimensions of the products to be analyzed, where the current indicator set of each dimension includes multiple current indicators. Then, the current keywords in the current indicators in the current indicator set of each dimension are extracted, and the current importance corresponding to the current indicator is determined based on the current keywords. Based on the current importance, the current weight corresponding to the current indicator is determined. Then, the indicator data of the current indicator is obtained, and the indicator data is input into the trained evaluation value prediction model to obtain the predicted evaluation value corresponding to the indicator data. The predicted evaluation value is fused according to the current weight to obtain the fused evaluation value corresponding to the current indicator. Then, based on the fused evaluation value, the current product classification result of the product to be analyzed is determined. Finally, the product demand information of the object to be analyzed is obtained, and the products to be analyzed whose current product classification results meet the product demand information are screened from each product to be analyzed as the target product corresponding to the object to be analyzed, and the target product is recommended to the object to be analyzed. In this way, in the process of recommending products, by extracting the current keywords in the current indicators in the current indicator set of each dimension of the product to be analyzed, determining the current weight corresponding to the current indicator, and combining the trained evaluation value prediction model to predict the indicator data of the current indicator, the fusion evaluation value corresponding to the current indicator can be determined more accurately, thereby more accurately determining the current product classification result of the product to be analyzed, and then based on the product demand information of the object to be analyzed, the target product corresponding to the product demand information can be more accurately screened from the product to be analyzed and recommended to the object to be analyzed, which is conducive to improving the accuracy of product recommendations; moreover, the entire process does not require human intervention, avoiding the subjective factors of manual judgment, which is prone to errors and leads to the defect of low accuracy of product recommendations, further improving the accuracy of product recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments of the present application or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0064] Figure 1 A flowchart of a model-based strategic emerging industry product recommendation method in one embodiment;

[0065] Figure 2 is a schematic diagram of a classification model in one embodiment;

[0066] Figure 3 is a schematic diagram of a classification model in another embodiment;

[0067] Figure 4 A flowchart of a model-based strategic emerging industry product recommendation method in another embodiment;

[0068] Figure 5 A structural block diagram of a model-based strategic emerging industry product recommendation device in one embodiment;

[0069] Figure 6 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0070] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0071] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0072] In an exemplary embodiment, Figure 1 As shown, a model-based strategic emerging industry product recommendation method is provided. This embodiment uses the method applied to a server as an example for illustration; it is understandable that the method can also be applied to a terminal, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, laptops, smart phones, and tablet computers; the server can be implemented as an independent server or a server cluster consisting of multiple servers. In this embodiment, the method includes the following steps:

[0073] Step S101, determining a current indicator set of multiple dimensions of a product to be analyzed; the product to be analyzed is used to represent the output of a strategic emerging industry; the current indicator set of each dimension includes multiple current indicators.

[0074] Among them, the products to be analyzed are used to represent the products output by strategic emerging industries (including goods, services and engineering).

[0075] The current indicator set for each dimension includes multiple current indicators. It should be noted that each dimension here consists of three dimensions: technological innovation, industrial control, and security support, as shown in Table 1. The technological innovation dimension includes current indicators such as indicator advancement, advanced functions, technological potential, R&D difficulty, and criticality of technical difficulties; the industrial control dimension includes current indicators such as policy compliance, digitalization, green development, strategic relevance, industry leadership, and market competitiveness; and the security support dimension includes current indicators such as risk impact, autonomous control capabilities, and supply chain resilience.

[0076] Table 1 Summary of product dimension indicator weights

[0077]

[0078] Exemplarily, the server obtains the performance indicators and technical parameters of the candidate products, and determines the performance indicator thresholds and technical parameter thresholds associated with the strategic emerging industries; then, the server selects the candidate products whose performance indicators meet the performance indicator thresholds and whose technical parameters meet the technical parameter thresholds from each candidate product, as products produced by the strategic emerging industries; then, the server uses the products produced by the strategic emerging industries as the products to be analyzed; then, the server determines the current indicators of the products to be analyzed, and based on these current indicators and the dimensions corresponding to these current indicators, constructs a current indicator set of multiple dimensions of the products to be analyzed.

[0079] Step S102 : extract the current keyword in the current indicator in the current indicator set of each dimension, and determine the current importance corresponding to the current indicator according to the current keyword.

[0080] Among them, the current keyword refers to the keyword in the current indicator, for example, the keyword in the current indicator "indicator advancement" is "indicator".

[0081] Among them, the current importance is used to indicate the importance corresponding to the current indicator.

[0082] Exemplarily, the server performs word segmentation processing on the current indicator in the current indicator set of each dimension to obtain multiple words corresponding to the current indicator; then, the server performs feature extraction processing on each word to obtain the word vector of each word, and inputs the word vector of each word into the importance prediction model to obtain the importance of each word; then, the server selects the word with the greatest importance from each word as the current keyword in the current indicator in the current indicator set of each dimension; then, the server inputs the current keyword into the hierarchical analysis model to obtain the current importance corresponding to the current indicator.

[0083] Step S103: Determine the current weight corresponding to the current indicator according to the current importance.

[0084] Among them, the current weight refers to the weight corresponding to the current indicator. As shown in Table 1, the predicted evaluation value corresponding to the current indicator "Indicator Advancement" is A1.

[0085] Exemplarily, the server queries the correspondence between importance and weight according to the current importance, obtains the weight corresponding to the current importance, and uses it as the current weight corresponding to the current indicator.

[0086] Step S104, obtain the indicator data of the current indicator, input the indicator data into the trained evaluation value prediction model, obtain the predicted evaluation value corresponding to the indicator data, and fuse the predicted evaluation value according to the current weight to obtain the fused evaluation value corresponding to the current indicator.

[0087] Among them, indicator data refers to the factor data corresponding to the current indicator. For example, the indicator data corresponding to the current indicator "indicator advancement" is speed, accuracy, energy consumption, capacity, etc., and the indicator data corresponding to the current indicator "functional advancement" is new functions that do not yet exist on the market.

[0088] Among them, the evaluation value prediction model refers to a network model that can use indicator data to obtain the predicted evaluation value corresponding to the indicator data, such as a convolutional neural network model and a recurrent neural network model.

[0089] It should be noted that the evaluation value prediction model corresponds to three different dimensions (technological innovation dimension, industrial control dimension and security support dimension), such as Figure 2 As shown. Existing models (such as the Kraljic model) have two corresponding dimensions (value impact and supply risk), as shown in Figure 3 shown.

[0090] The predicted evaluation value is used to represent the score corresponding to the indicator data. As shown in Table 1, the predicted evaluation value corresponding to the current indicator "Indicator Advancement" is X1.

[0091] The fusion evaluation value refers to the evaluation value obtained by fusing the predicted evaluation values according to the current weight. As shown in Table 1, the fusion evaluation value corresponding to the current indicator "Indicator Advancement" is A1×X1, and the fusion evaluation value corresponding to the current indicator "Function Advancement" is B1×X2.

[0092] Exemplarily, the server obtains the indicator data of the current indicator from the database; then, the server preprocesses the indicator data to obtain preprocessed indicator data, and performs feature extraction on the preprocessed indicator data to obtain a feature vector of the preprocessed indicator data; then, the server inputs the feature vector of the preprocessed indicator data into the trained evaluation value prediction model to obtain a first evaluation value corresponding to the indicator data, and inputs the feature vector of the preprocessed indicator data into the historical evaluation value prediction model corresponding to the trained evaluation value prediction model to obtain a second evaluation value corresponding to the indicator data; then, the server determines the first model weight of the trained evaluation value prediction model and the second model weight of the historical evaluation value prediction model, and sums the first evaluation value and the second evaluation value according to the first model weight and the second model weight to obtain a predicted evaluation value corresponding to the indicator data; then, the server multiplies the predicted evaluation value according to the current weight to obtain a fusion evaluation value corresponding to the current indicator.

[0093] Step S105: Determine the current product classification result of the product to be analyzed based on the fusion evaluation value.

[0094] Among them, the current product classification results refer to the product classification results of the products to be analyzed, including strategic core products, innovation-leading products, safety-critical products, cutting-edge exploration products, traditional pillar products, scale advantage products, basic guarantee products and ordinary products.

[0095] Exemplarily, the server determines the product classification result of the product to be analyzed in each dimension based on the fusion evaluation value; then, the server determines the current product classification result of the product to be analyzed based on the product classification result of the product to be analyzed in each dimension.

[0096] Step S106 , obtaining product demand information of the object to be analyzed, screening out products to be analyzed whose current product classification results meet the product demand information from the products to be analyzed, taking them as target products corresponding to the object to be analyzed, and recommending the target products to the object to be analyzed.

[0097] The object to be analyzed refers to the subject who has product demand or needs to be recommended products, such as users.

[0098] The product demand information is used to indicate the specific requirements of the object to be analyzed for the required products.

[0099] The target product is used to represent the product that matches the object to be analyzed.

[0100] Exemplarily, the server obtains the product purchase record information and historical browsing information of the object to be analyzed, and inputs the product purchase record information and historical browsing information of the object to be analyzed into the trained product demand prediction model to obtain the product demand information of the object to be analyzed; then, the server filters out the products to be analyzed whose current product classification results meet the product demand information from each product to be analyzed, as the target products corresponding to the object to be analyzed; then, the server generates product recommendation information corresponding to the target product, and sends the product recommendation information to the terminal corresponding to the object to be analyzed.

[0101] In the above-mentioned model-based strategic emerging industry product recommendation method, the products produced by strategic emerging industries are first taken as products to be analyzed, and the current indicator sets of multiple dimensions of the products to be analyzed are determined, and the current indicator set of each dimension includes multiple current indicators. Then, the current keywords in the current indicators in the current indicator set of each dimension are extracted, and the current importance corresponding to the current indicator is determined according to the current keywords, and the current weight corresponding to the current indicator is determined according to the current importance. Then, the indicator data of the current indicator is obtained, and the indicator data is input into the trained evaluation value prediction model to obtain the predicted evaluation value corresponding to the indicator data, and the predicted evaluation value is fused according to the current weight to obtain the fused evaluation value corresponding to the current indicator. Then, according to the fused evaluation value, the current product classification result of the product to be analyzed is determined. Finally, the product demand information of the object to be analyzed is obtained, and the products to be analyzed whose current product classification results meet the product demand information are screened out from each product to be analyzed as the target product corresponding to the object to be analyzed, and the target product is recommended to the object to be analyzed. In this way, in the process of recommending products, by extracting the current keywords in the current indicators in the current indicator set of each dimension of the product to be analyzed, determining the current weight corresponding to the current indicator, and combining the trained evaluation value prediction model to predict the indicator data of the current indicator, the fusion evaluation value corresponding to the current indicator can be determined more accurately, thereby more accurately determining the current product classification result of the product to be analyzed, and then based on the product demand information of the object to be analyzed, the target product corresponding to the product demand information can be more accurately screened from the product to be analyzed and recommended to the object to be analyzed, which is conducive to improving the accuracy of product recommendations; moreover, the entire process does not require human intervention, avoiding the subjective factors of manual judgment, which is prone to errors and leads to the defect of low accuracy of product recommendations, further improving the accuracy of product recommendations.

[0102] In an exemplary embodiment, the above step S102 determines the current importance corresponding to the current indicator based on the current keyword, specifically including the following contents: obtaining similar keywords corresponding to the current keyword; determining the first importance of the current keyword based on the current keyword, similar keywords and the long text associated with the product to be analyzed; determining the second importance of the current keyword based on the current keyword, similar keywords and the trained sentiment prediction model; determining the third importance of the current keyword based on the preset correspondence between the current keyword, similar keywords and pre-constructed keywords and importance; fusing the first importance, the second importance and the third importance to obtain the current importance.

[0103] Among them, similar keywords refer to keywords whose similarity with the current keyword is greater than a preset similarity.

[0104] The long texts associated with the product to be analyzed include industry reports and analysis articles in the field of the product to be analyzed.

[0105] The first importance refers to the importance of the current keyword determined based on the current keyword, similar keywords, and long text associated with the product to be analyzed.

[0106] Among them, the sentiment prediction model refers to a network model that can predict the sentiment tendency of keywords, such as the Naive Bayes model.

[0107] The second importance refers to the importance of the current keyword determined based on the current keyword, similar keywords and the trained sentiment prediction model.

[0108] Among them, the preset corresponding relationship is used to represent the association information between keywords and importance. For example, the keyword "indicator" in "indicator advancement" corresponds to an importance of 0.6, and the keyword "function" in "function advancement" corresponds to an importance of 0.7.

[0109] The third importance refers to the importance of the current keyword determined based on the preset correspondence between the current keyword, similar keywords, and pre-established keywords and importances.

[0110] Exemplarily, the server inputs the current keyword into the trained text generation model to obtain candidate keywords related to the current keyword; then, the server performs feature extraction processing on the current keyword and the candidate keyword respectively to obtain the first word vector of the current keyword and the second word vector of the candidate keyword, and determines the first similarity between the current keyword and the candidate keyword based on the difference between the first word vector and the second word vector; then, the server determines the current edit distance between the current keyword and the candidate keyword, and queries the corresponding relationship between the edit distance and the similarity to obtain the second similarity between the current keyword and the candidate keyword; then, the server performs fusion processing on the first similarity and the second similarity to obtain the first similarity between the current keyword and the candidate keyword. The target similarity between keywords is determined, and candidate keywords with target similarity greater than the preset similarity are screened out from each candidate keyword as similar keywords corresponding to the current keyword; then, the server determines the first importance of the current keyword based on the current keyword, similar keywords and the long text associated with the product to be analyzed, determines the second importance of the current keyword based on the current keyword, similar keywords and the trained sentiment prediction model, and determines the third importance of the current keyword based on the preset correspondence between the current keyword, similar keywords and pre-built keywords and importance; then, the server sums and averages the first importance, the second importance and the third importance to obtain the current importance corresponding to the current indicator.

[0111] In this embodiment, obtaining similar keywords to the current keyword can consider the problem from a broader semantic perspective. Similar keywords can help explore different expressions and concepts related to the current keyword, enrich the understanding of the meaning of the keyword, avoid the limitations of analyzing only from a single vocabulary perspective, and help improve the accuracy of determining the current importance corresponding to the current indicator.

[0112] In an exemplary embodiment, the first importance of the current keyword is determined based on the current keyword, similar keywords and long text associated with the product to be analyzed, which specifically also includes the following contents: determining the first occurrence frequency of the current keyword in the long text, and the second occurrence frequency of similar keywords in the long text; fusing the first occurrence frequency and the second occurrence frequency to obtain the target occurrence frequency corresponding to the current keyword; and determining the first importance based on the target occurrence frequency.

[0113] The first occurrence frequency refers to the occurrence frequency of the current keyword in the long text.

[0114] The second occurrence frequency refers to the frequency of occurrence of similar keywords in the long text.

[0115] The target occurrence frequency refers to an occurrence frequency obtained by fusing the first occurrence frequency and the second occurrence frequency.

[0116] Exemplarily, the server determines the first number of occurrences of the current keyword in the long text, and the second number of occurrences of similar keywords in the long text; then, the server obtains the number of words in the long text, and determines the first frequency of occurrence of the current keyword in the long text based on the first number of occurrences and the number of words, and determines the second frequency of occurrence of similar keywords in the long text based on the second number of occurrences and the number of words; then, the server determines the first occurrence position of the current keyword in the long text, and the second occurrence position of the similar keyword in the long text, and queries the correspondence between the occurrence position and the preset weight to obtain the first preset weight corresponding to the first occurrence frequency, and the second preset weight corresponding to the second occurrence frequency; then, the server sums the first occurrence frequency and the second occurrence frequency according to the first preset weight and the second preset weight to obtain the target occurrence frequency corresponding to the current keyword; then, the server queries the correspondence between the occurrence frequency and the importance based on the target occurrence frequency as the first importance of the current keyword.

[0117] In this embodiment, by comprehensively considering the frequency of occurrence of the current keyword and similar keywords, the importance of the current keyword can be supplemented from the side. By fusing the first frequency of occurrence and the second frequency of occurrence, the target frequency of occurrence obtained can more comprehensively reflect the importance of the current keyword in the long text, avoiding the one-sidedness of judging the importance based solely on the frequency of occurrence of the current keyword.

[0118] In an exemplary embodiment, the second importance of the current keyword is determined based on the current keyword, similar keywords and the trained emotion prediction model, which specifically also includes the following contents: the current keyword and similar keywords are respectively input into the trained emotion prediction model to obtain the first emotion prediction result of the current keyword and the second emotion prediction result of the similar keyword; the correspondence between the emotion prediction results and the importance is queried to obtain the fourth importance corresponding to the first emotion prediction result and the fifth importance corresponding to the second emotion prediction result; the first weight corresponding to the first emotion prediction result is determined based on the first prediction probability corresponding to the first emotion prediction result output by the trained emotion prediction model, and the second weight corresponding to the second emotion prediction result is determined based on the second prediction probability corresponding to the second emotion prediction result output by the trained emotion prediction model; the fourth importance and the fifth importance are summed according to the first weight and the second weight to obtain the second importance.

[0119] The first emotion prediction result is used to represent the emotion prediction result of the current keyword.

[0120] The second emotion prediction result is used to represent the emotion prediction result of similar keywords.

[0121] Among them, the fourth importance is used to represent the importance corresponding to the first emotion prediction result.

[0122] Among them, the fifth importance is used to represent the importance corresponding to the second emotion prediction result.

[0123] Among them, the first prediction probability is used to represent the prediction probability corresponding to the first emotion prediction result.

[0124] Among them, the first weight is used to represent the weight corresponding to the first emotion prediction result.

[0125] Among them, the second prediction probability is used to represent the prediction probability corresponding to the second emotion prediction result.

[0126] The second weight is used to represent the weight corresponding to the second emotion prediction result.

[0127] Exemplarily, the server inputs the current keyword into the trained emotion prediction model to obtain the prediction probability of the current keyword under each first preset emotion prediction result, and screens out the first preset emotion prediction result with the largest prediction probability from each first preset emotion prediction result as the first emotion prediction result of the current keyword; then, the server inputs similar keywords into the trained emotion prediction model to obtain the prediction probability of similar keywords under each second preset emotion prediction result, and screens out the second preset emotion prediction result with the largest prediction probability from each second preset emotion prediction result as the second emotion prediction result of the similar keyword; then, the server queries the correspondence between the emotion prediction result and the importance based on the first emotion prediction result and the second emotion prediction result. The fourth importance corresponding to the first emotion prediction result and the fifth importance corresponding to the second emotion prediction result are obtained; then, the server queries the correspondence between the prediction probability and the weight according to the first prediction probability corresponding to the first emotion prediction result output by the trained emotion prediction model, and obtains the weight corresponding to the first prediction probability as the first weight corresponding to the first emotion prediction result; then, the server queries the correspondence between the prediction probability and the weight according to the second prediction probability corresponding to the second emotion prediction result output by the trained emotion prediction model, and obtains the weight corresponding to the second prediction probability as the second weight corresponding to the second emotion prediction result; then, the server sums the fourth importance and the fifth importance according to the first weight and the second weight to obtain the second importance.

[0128] In this embodiment, by inputting the current keyword and similar keywords into the sentiment prediction model, the sentiment factor is introduced to measure the importance of the keyword, and by querying the correspondence between the sentiment prediction result and the importance, the sentiment prediction result is converted into a specific importance value, so that the influence of the sentiment factor on the importance can be quantified.

[0129] In an exemplary embodiment, the third importance of the current keyword is determined based on the preset correspondence between the current keyword, similar keywords, and pre-constructed keywords and importance, which specifically also includes the following contents: querying the preset correspondence to obtain the sixth importance corresponding to the current keyword and the seventh importance corresponding to the similar keyword; determining the preset weight corresponding to the similar keyword based on the similarity between the current keyword and the similar keyword; and summing the sixth importance and the seventh importance according to the basic weight and the preset weight of the current keyword to obtain the third importance.

[0130] The sixth importance is used to indicate the importance corresponding to the current keyword.

[0131] The seventh importance is used to indicate the importance corresponding to similar keywords.

[0132] The similarity between the current keyword and the similar keywords is used to indicate the degree of similarity between the current keyword and the similar keywords.

[0133] The preset weight is determined by the similarity between the current keyword and similar keywords. In actual scenarios, the greater the similarity between the current keyword and similar keywords, the greater the preset weight corresponding to the similar keywords. For example, if the similarity between the current keyword and similar keywords is greater than the preset similarity, the preset weight corresponding to the similar keywords is 0.9; if the similarity between the current keyword and similar keywords is less than or equal to the preset similarity, the preset weight corresponding to the similar keywords is 0.7.

[0134] Among them, the basic weight is used to represent the initial weight of the current keyword, such as 1.

[0135] Exemplarily, the server queries the preset correspondence between pre-constructed keywords and importance, and obtains the sixth importance corresponding to the current keyword, and the seventh importance corresponding to the similar keyword; then, the server queries the correspondence between similarity and weight based on the similarity between the current keyword and the similar keyword, and obtains the preset weight corresponding to the similar keyword; then, the server obtains the basic weight of the current keyword, and sums the sixth importance and the seventh importance according to the basic weight and the preset weight of the current keyword to obtain the third importance.

[0136] In this embodiment, the importance of the current keyword and similar keywords is obtained by querying the preset correspondence, and the basic weight of the current keyword and the preset weight of the similar keyword determined according to the similarity between the two are taken into consideration. This can more comprehensively and carefully evaluate the importance of the keyword in the text, avoiding the one-sidedness of measuring the importance of the keyword from only a single perspective.

[0137] In an exemplary embodiment, the current indicator sets of multiple dimensions include a first indicator set, a second indicator set, and a third indicator set.

[0138] Then, the above-mentioned step S105 determines the current product classification result of the product to be analyzed based on the fusion evaluation value, and specifically also includes the following contents: according to the fusion evaluation value, determine the first evaluation value of the product to be analyzed under the first indicator set, the second evaluation value of the product to be analyzed under the second indicator set, and the third evaluation value of the product to be analyzed under the third indicator set; according to the first evaluation value, the second evaluation value and the third evaluation value, query the correspondence between the evaluation value and the product classification result, and obtain the first product classification result of the product to be analyzed under the first indicator set, the second product classification result of the product to be analyzed under the second indicator set, and the third product classification result of the product to be analyzed under the third indicator set; obtain the product category corresponding to the first product classification result, the second product classification result and the third product classification result as the current product classification result.

[0139] Among them, the first indicator set refers to the indicator set of the scientific and technological innovation dimension.

[0140] Among them, the second indicator set refers to the indicator set of the industrial control dimension.

[0141] Among them, the third indicator set refers to the indicator set of the security support dimension.

[0142] The first evaluation value refers to the evaluation value under the first indicator set.

[0143] Referring to Table 1, the first evaluation value is also called the comprehensive score of scientific and technological innovation, which can be calculated using the following formula:

[0144] Comprehensive score of scientific and technological innovation = A1×X1+B1×X2+C1×X3+D1×X4+E1×X5, formula (1)

[0145] The second evaluation value refers to the evaluation value under the second indicator set.

[0146] Referring to Table 1, the second evaluation value is also called the comprehensive score of industrial control, which can be calculated using the following formula:

[0147] Comprehensive score of industrial control = A2×Y1+B2×Y2+C2×Y3+D2×Y4+E2×Y5+F2×Y6, formula (2)

[0148] The third evaluation value refers to the evaluation value under the third indicator set.

[0149] Referring to Table 1, the third evaluation value is also called the safety support comprehensive score, which can be calculated using the following formula:

[0150] Safety support comprehensive score = A3×Z1+B3×Z2+C3×Z3, formula (3)

[0151] The corresponding relationship between the evaluation value and the product classification result is used to represent the association information between the evaluation value and the product classification result. For example, when the evaluation value is greater than 50, the corresponding product classification result is high; when the evaluation value is less than or equal to 50, the corresponding product classification result is low.

[0152] The first product classification result refers to the product classification result of the product to be analyzed under the first indicator set, including high and low.

[0153] The second product classification result refers to the product classification result of the product to be analyzed under the second indicator set, including high and low.

[0154] The third product classification result refers to the product classification result of the product to be analyzed under the third indicator set, including high and low.

[0155] Among them, product category refers to the category to which the product to be analyzed belongs, including strategic core products, innovation-leading products, safety-critical products, cutting-edge exploration products, traditional pillar products, scale advantage products, basic guarantee products and ordinary products.

[0156] Referring to Table 2, when the first product classification result is high, the second product classification result is high, and the third product classification result is high, the corresponding product category is strategic core product; when the first product classification result is high, the second product classification result is high, and the third product classification result is low, the corresponding product category is innovation-led product; when the first product classification result is high, the second product classification result is low, and the third product classification result is high, the corresponding product category is safety-critical product; when the first product classification result is high, the second product classification result is low, and the third product classification result is low, the corresponding product category is frontier exploration product; when the first product classification result is low, the second product classification result is high, and the third product classification result is high, the corresponding product category is traditional pillar product; when the first product classification result is low, the second product classification result is high, and the third product classification result is low, the corresponding product category is scale advantage product; when the first product classification result is low, the second product classification result is low, and the third product classification result is high, the corresponding product category is basic support product; when the first product classification result is low, the second product classification result is low, and the third product classification result is low, the corresponding product category is ordinary product.

[0157] Table 2 Product classification results

[0158]

[0159] Exemplarily, the server extracts a first sub-fusion evaluation value corresponding to the first indicator set from the fusion evaluation value, and sums the first sub-fusion evaluation values corresponding to the first indicator set to obtain a first evaluation value of the product to be analyzed under the first indicator set; then, the server extracts a second sub-fusion evaluation value corresponding to the second indicator set from the fusion evaluation value, and sums the second sub-fusion evaluation values corresponding to the second indicator set to obtain a second evaluation value of the product to be analyzed under the second indicator set; then, the server extracts a third sub-fusion evaluation value corresponding to the third indicator set from the fusion evaluation value, and sums the third sub-fusion evaluation value corresponding to the third indicator set to obtain a third evaluation value of the product to be analyzed under the third indicator set; then, according to the first evaluation value, the second evaluation value and the third evaluation value, query the relationship between the evaluation value and the product classification result The server constructs a correspondence between the first product classification result, the second product classification result, and the third product classification result, and obtains the product classification result corresponding to the first evaluation value, the product classification result corresponding to the second evaluation value, and the product classification result corresponding to the third evaluation value, respectively as the first product classification result of the product to be analyzed under the first indicator set, the second product classification result of the product to be analyzed under the second indicator set, and the third product classification result of the product to be analyzed under the third indicator set; then, the server constructs a correspondence between the first product classification result, the second product classification result, the third product classification result and the product category as the target correspondence; then, the server queries the target correspondence according to the first product classification result, the second product classification result, and the third product classification result, and obtains the product category corresponding to the first product classification result, the second product classification result, and the third product classification result, and uses the product category as the current product classification result.

[0160] In this embodiment, by determining the evaluation value of the product to be analyzed under different sets of indicators, the product can be comprehensively considered from multiple dimensions, thereby avoiding evaluating the product based on only a single indicator, making the evaluation results more comprehensive, objective and accurate in reflecting the true situation of the product, which is conducive to improving the accuracy of determining the current product classification results.

[0161] In an exemplary embodiment, Figure 4 As shown, another model-based strategic emerging industry product recommendation method is provided. This method is applied to a server as an example and specifically includes the following steps:

[0162] Step S401, determining a current indicator set of multiple dimensions of a product to be analyzed; the product to be analyzed is used to represent the output of a strategic emerging industry; the current indicator set of each dimension includes multiple current indicators.

[0163] Step S402: extract the current keyword in the current indicator in the current indicator set of each dimension.

[0164] Step S403: Obtain similar keywords corresponding to the current keyword.

[0165] Step S404: determine the first frequency of occurrence of the current keyword in the long text, and the second frequency of occurrence of similar keywords in the long text; fuse the first frequency of occurrence and the second frequency of occurrence to obtain a target frequency of occurrence corresponding to the current keyword; and determine the first importance based on the target frequency of occurrence.

[0166] In step S405, the current keyword and similar keywords are respectively input into the trained emotion prediction model to obtain the first emotion prediction result of the current keyword and the second emotion prediction result of the similar keyword; the correspondence between the emotion prediction results and the importance is queried to obtain the fourth importance corresponding to the first emotion prediction result and the fifth importance corresponding to the second emotion prediction result.

[0167] Step S406: Determine the first weight corresponding to the first emotion prediction result based on the first prediction probability corresponding to the first emotion prediction result output by the trained emotion prediction model, and determine the second weight corresponding to the second emotion prediction result based on the second prediction probability corresponding to the second emotion prediction result output by the trained emotion prediction model.

[0168] Step S407 : According to the first weight and the second weight, the fourth importance and the fifth importance are summed to obtain the second importance.

[0169] Step S408 , querying the preset corresponding relationship, obtaining the sixth importance corresponding to the current keyword, and the seventh importance corresponding to the similar keyword; and determining the preset weight corresponding to the similar keyword based on the similarity between the current keyword and the similar keyword.

[0170] Step S409 : According to the basic weight and the preset weight of the current keyword, the sixth importance and the seventh importance are summed to obtain the third importance.

[0171] Step S410 , fusing the first importance, the second importance, and the third importance to obtain the current importance.

[0172] Step S411: Determine the current weight corresponding to the current indicator according to the current importance.

[0173] Step S412, obtain the indicator data of the current indicator, input the indicator data into the trained evaluation value prediction model, obtain the predicted evaluation value corresponding to the indicator data, and fuse the predicted evaluation value according to the current weight to obtain the fused evaluation value corresponding to the current indicator.

[0174] Step S413: Determine the current product classification result of the product to be analyzed based on the fusion evaluation value.

[0175] Step S414 , obtaining product demand information of the object to be analyzed, screening out products to be analyzed whose current product classification results meet the product demand information from the products to be analyzed, taking them as target products corresponding to the object to be analyzed, and recommending the target products to the object to be analyzed.

[0176] In the above-mentioned model-based strategic emerging industry product recommendation method, in the process of recommending products, by extracting the current keywords in the current indicators in the current indicator set of each dimension of the product to be analyzed, the current weight corresponding to the current indicator is determined, and the indicator data of the current indicator is predicted in combination with the trained evaluation value prediction model, the fusion evaluation value corresponding to the current indicator can be determined more accurately, so that the current product classification result of the product to be analyzed can be determined more accurately, and then based on the product demand information of the object to be analyzed, the target product corresponding to the product demand information can be more accurately screened from the product to be analyzed and recommended to the object to be analyzed, which is conducive to improving the accuracy of product recommendation; moreover, the entire process does not require human intervention, avoiding the defects of subjective factors and easy errors in the manual judgment method, resulting in low accuracy of product recommendation, and further improving the accuracy of product recommendation.

[0177] In an exemplary embodiment, in order to more clearly illustrate the model-based strategic emerging industry product recommendation method provided in the embodiment of this application, the following specific embodiment is used to specifically illustrate the model-based strategic emerging industry product recommendation method. In one embodiment, this application also provides a "three forces" classification model for strategic emerging industry products. Specifically, it includes the following content:

[0178] The "three forces" classification model for strategic emerging industry products consists of three dimensions: scientific and technological innovation, industrial control, and security support. Each dimension is divided into two intervals of high and low, and strategic emerging industry products can be divided into 8 intervals.

[0179] According to the classification principles, new products can be divided into eight categories, including strategic core products, innovation-leading products, safety-critical products, cutting-edge exploration products, traditional pillar products, scale advantage products, basic support products, and ordinary products. The specific classification is shown in Table 2.

[0180] Evaluation indicators are designed for each dimension, and the range is judged scientifically.

[0181] The technological innovation dimension primarily considers five indicators: advanced indicators, advanced functions, technological potential, R&D difficulty, and criticality of technical difficulties. The industrial control dimension primarily considers policy compliance, digitalization, green development, strategic relevance, industry leadership, and market competitiveness. The security support dimension considers products based on three indicators: risk impact, autonomous control capabilities, and supply chain resilience.

[0182] Using the analytic hierarchy process, we calculated the weight of each indicator in each dimension through importance comparison, and then used the comprehensive evaluation method and expert scoring to finally determine the product classification range. The calculation process is as follows:

[0183] Invite 5-9 (odd number) procurement experts, industry experts, and internal and external experts in emerging industries to compare indicators at the same level (grade) based on their relative importance and generate scores. The weighting process is computer-assisted and the steps are as follows:

[0184] Step 1: Build a hierarchical model.

[0185] Based on the characteristics and objectives of the problem, the problem is broken down into different levels, generally including the goal level, the criterion level and the solution level.

[0186] Step 2: Construct a judgment matrix.

[0187] For each element at the same level, the relative importance of each element is compared pairwise to construct a judgment matrix. The 1-9 scale is used for comparison.

[0188] The comparative judgment matrix is quantified using a nine-level scale method, and the scale of 1-9 can reflect the judgment ability of most people.

[0189] Step 3: Calculate the weight vector.

[0190] Solve the maximum eigenvalue and corresponding eigenvector of the judgment matrix, normalize the eigenvector, and obtain the weight of each element.

[0191] Calculation formula:

[0192] , formula (4)

[0193] For each comparison judgment matrix A, the maximum eigenvalue λmax and eigenvector α of the comparison judgment matrix A satisfying the equation Aα=λmaxα are calculated, where α is the weight vector, that is, the weight of each indicator.

[0194] Step 4: Index consistency test.

[0195] In order to ensure the consistency of the judgments when judging the relative importance of each factor and avoid contradictions, after finding λmax, it is necessary to test the consistency of the comparison judgment matrix. The consistency index is calculated according to the formula:

[0196] , formula (5)

[0197] Among them, n is the order of the comparison judgment matrix;

[0198] The consistency ratio is calculated according to the formula:

[0199] CR=CI / RI, formula (6)

[0200] Among them, RI is the average consistency index, which can be obtained by looking up the table:

[0201] It is generally believed that when CR < 0.10, the judgment matrix has good consistency. Otherwise, the judgment matrix needs to be adjusted to meet the consistency test.

[0202] (5) After calculating the weights of the evaluation indicators for each dimension of technological innovation, industrial control, and security support, we compiled the weights of the "three forces" dimension indicators for new products in Table 1. We invited experts to score the new products based on the three dimensions, using a 100-point scale. Based on the weights in Table 1, we multiplied the score by the weight and added them together to obtain the comprehensive score for each dimension. Based on the comprehensive score, we divided the new products into different categories.

[0203] Comprehensive score of scientific and technological innovation = A1×X1+B1×X2+C1×X3+D1×X4+E1×X5, formula (1)

[0204] Comprehensive score of industrial control = A2×Y1+B2×Y2+C2×Y3+D2×Y4+E2×Y5+F2×Y6, formula (2)

[0205] Safety support comprehensive score = A3×Z1+B3×Z2+C3×Z3, formula (3)

[0206] Collect and summarize expert ratings, and use the average value to calculate the score of each product after the "three forces" evaluation to determine the product range.

[0207] In the above embodiment, during the product recommendation process, by extracting the current keywords from the current indicators in the current indicator set for each dimension of the product to be analyzed, determining the current weight corresponding to the current indicator, and predicting the indicator data of the current indicator in combination with the trained evaluation value prediction model, the fused evaluation value corresponding to the current indicator can be more accurately determined, thereby more accurately determining the current product classification result of the product to be analyzed. Furthermore, based on the product demand information of the object to be analyzed, target products corresponding to the product demand information can be more accurately screened from the products to be analyzed and recommended to the object to be analyzed, which is conducive to improving the accuracy of product recommendations. Moreover, the entire process does not require human intervention, avoiding the subjective factors and prone to errors in manual judgment, which leads to low product recommendation accuracy, further improving the accuracy of product recommendations. At the same time, the "three forces" model of new strategic products extends the traditional two-dimensional classification to three-dimensional classification, combines the characteristics of new strategic products, and can more accurately classify new strategic products, thereby helping to more accurately formulate differentiated procurement strategies for each range of new strategic products, and use procurement to drive the development of the new strategic industry. Moreover, the introduction of computer programs to assist in the classification process can effectively improve the efficiency of product classification and enhance scientificity.

[0208] It should be understood that, although the steps in the flowcharts of the above embodiments are shown in sequence as indicated by the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be performed in other orders. Moreover, at least a portion of the steps in the flowcharts of the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily to be performed in sequence, but can be performed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0209] Based on the same inventive concept, the embodiments of the present application also provide a model-based strategic emerging industry product recommendation device for implementing the aforementioned model-based strategic emerging industry product recommendation method. The implementation solution provided by the device is similar to the implementation solution described in the aforementioned method. Therefore, the specific limitations in one or more embodiments of the model-based strategic emerging industry product recommendation device provided below can be found in the limitations of the model-based strategic emerging industry product recommendation method above, and will not be repeated here.

[0210] In an exemplary embodiment, Figure 5As shown, a model-based strategic emerging industry product recommendation device is provided, comprising: a set determination module 501, an importance determination module 502, a weight determination module 503, an evaluation value determination module 504, a result determination module 505, and a product recommendation module 506, wherein:

[0211] The set determination module 501 is used to determine the current indicator sets of multiple dimensions of the product to be analyzed; the product to be analyzed is used to represent the products output by strategic emerging industries; the current indicator set of each dimension includes multiple current indicators.

[0212] The importance determination module 502 is used to extract the current keyword in the current indicator in the current indicator set of each dimension, and determine the current importance corresponding to the current indicator according to the current keyword.

[0213] The weight determination module 503 is used to determine the current weight corresponding to the current indicator according to the current importance.

[0214] The evaluation value determination module 504 is used to obtain the indicator data of the current indicator, input the indicator data into the trained evaluation value prediction model, obtain the predicted evaluation value corresponding to the indicator data, and fuse the predicted evaluation value according to the current weight to obtain the fused evaluation value corresponding to the current indicator.

[0215] The result determination module 505 is used to determine the current product classification result of the product to be analyzed according to the fusion evaluation value.

[0216] The product recommendation module 506 is used to obtain product demand information of the object to be analyzed, filter out the products to be analyzed whose current product classification results meet the product demand information from each product to be analyzed, use them as target products corresponding to the object to be analyzed, and recommend the target products to the object to be analyzed.

[0217] In an exemplary embodiment, the importance determination module 502 is also used to obtain similar keywords corresponding to the current keyword; determine the first importance of the current keyword based on the current keyword, similar keywords and the long text associated with the product to be analyzed; determine the second importance of the current keyword based on the current keyword, similar keywords and the trained sentiment prediction model; determine the third importance of the current keyword based on the preset correspondence between the current keyword, similar keywords and pre-constructed keywords and importance; and fuse the first importance, the second importance and the third importance to obtain the current importance.

[0218] In an exemplary embodiment, the importance determination module 502 is also used to determine the first occurrence frequency of the current keyword in the long text, and the second occurrence frequency of similar keywords in the long text; the first occurrence frequency and the second occurrence frequency are merged to obtain the target occurrence frequency corresponding to the current keyword; and the first importance is determined based on the target occurrence frequency.

[0219] In an exemplary embodiment, the importance determination module 502 is also used to input the current keyword and similar keywords into the trained emotion prediction model respectively to obtain the first emotion prediction result of the current keyword and the second emotion prediction result of the similar keyword; query the correspondence between the emotion prediction results and the importance to obtain the fourth importance corresponding to the first emotion prediction result and the fifth importance corresponding to the second emotion prediction result; determine the first weight corresponding to the first emotion prediction result based on the first prediction probability corresponding to the first emotion prediction result output by the trained emotion prediction model, and determine the second weight corresponding to the second emotion prediction result based on the second prediction probability corresponding to the second emotion prediction result output by the trained emotion prediction model; sum the fourth importance and the fifth importance according to the first weight and the second weight to obtain the second importance.

[0220] In an exemplary embodiment, the importance determination module 502 is also used to query the preset correspondence to obtain the sixth importance corresponding to the current keyword and the seventh importance corresponding to the similar keyword; based on the similarity between the current keyword and the similar keyword, the preset weight corresponding to the similar keyword is determined; according to the basic weight and the preset weight of the current keyword, the sixth importance and the seventh importance are summed to obtain the third importance.

[0221] In an exemplary embodiment, the result determination module 505 is also used to determine, based on the fusion evaluation value, a first evaluation value of the product to be analyzed under the first indicator set, a second evaluation value of the product to be analyzed under the second indicator set, and a third evaluation value of the product to be analyzed under the third indicator set; based on the first evaluation value, the second evaluation value and the third evaluation value, query the correspondence between the evaluation value and the product classification result to obtain the first product classification result of the product to be analyzed under the first indicator set, the second product classification result of the product to be analyzed under the second indicator set, and the third product classification result of the product to be analyzed under the third indicator set; obtain the product category corresponding to the first product classification result, the second product classification result and the third product classification result as the current product classification result.

[0222] Each module in the aforementioned model-based strategic emerging industry product recommendation device can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0223] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 6 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O) and a communication interface. The processor, memory and input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store data such as current indicators and current keywords. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a model-based strategic emerging industry product recommendation method is implemented.

[0224] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0225] In an exemplary embodiment, a computer device is further provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0226] In an exemplary embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0227] In an exemplary embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0228] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), data processing logic devices based on quantum computing, and the like.

[0229] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0230] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A model-based method for recommending products in strategic emerging industries, characterized in that: The method comprises: Determining a current indicator set of multiple dimensions of a product to be analyzed; the product to be analyzed is used to represent a product output by a strategic emerging industry; the current indicator set of each dimension includes multiple current indicators; Extracting current keywords from the current indicators in the current indicator set of each dimension, and determining the current importance corresponding to the current indicator based on the current keywords; Determining a current weight corresponding to the current indicator according to the current importance; Obtaining indicator data of the current indicator, inputting the indicator data into a trained evaluation value prediction model to obtain a predicted evaluation value corresponding to the indicator data, and fusing the predicted evaluation values according to the current weight to obtain a fused evaluation value corresponding to the current indicator; Determining a current product classification result of the product to be analyzed based on the fusion evaluation value; Obtain product demand information of the object to be analyzed, filter out the products to be analyzed whose current product classification results meet the product demand information from the products to be analyzed, use them as target products corresponding to the object to be analyzed, and recommend the target products to the object to be analyzed.

2. The method according to claim 1, characterized in that Determining the current importance corresponding to the current indicator according to the current keyword includes: Obtain similar keywords corresponding to the current keyword; Determining a first importance of the current keyword based on the current keyword, the similar keywords, and the long text associated with the product to be analyzed; determining a second importance of the current keyword based on the current keyword, the similar keywords, and a trained sentiment prediction model; and determining a third importance of the current keyword based on a preset correspondence between the current keyword, the similar keywords, and pre-established keywords and importances; The first importance, the second importance, and the third importance are fused to obtain the current importance.

3. The method according to claim 2, characterized in that The determining of the first importance of the current keyword based on the current keyword, the similar keywords, and the long text associated with the product to be analyzed includes: Determining a first occurrence frequency of the current keyword in the long text and a second occurrence frequency of the similar keyword in the long text; performing a fusion process on the first occurrence frequency and the second occurrence frequency to obtain a target occurrence frequency corresponding to the current keyword; The first importance is determined according to the target occurrence frequency.

4. The method according to claim 2, characterized in that The determining of the second importance of the current keyword based on the current keyword, the similar keywords, and the trained sentiment prediction model includes: Inputting the current keyword and the similar keyword into the trained emotion prediction model respectively to obtain a first emotion prediction result of the current keyword and a second emotion prediction result of the similar keyword; Querying the correspondence between the emotion prediction results and the importance, and obtaining the fourth importance corresponding to the first emotion prediction result and the fifth importance corresponding to the second emotion prediction result; Determining a first weight corresponding to the first emotion prediction result based on a first prediction probability corresponding to the first emotion prediction result output by the trained emotion prediction model, and determining a second weight corresponding to the second emotion prediction result based on a second prediction probability corresponding to the second emotion prediction result output by the trained emotion prediction model; The fourth importance and the fifth importance are summed according to the first weight and the second weight to obtain the second importance.

5. The method according to claim 2, characterized in that The determining of the third importance of the current keyword according to the preset correspondence between the current keyword, the similar keywords, and the pre-established keywords and importances includes: Querying the preset corresponding relationship to obtain the sixth importance corresponding to the current keyword and the seventh importance corresponding to the similar keyword; Determining a preset weight corresponding to the similar keyword based on the similarity between the current keyword and the similar keyword; The sixth importance and the seventh importance are summed according to the basic weight of the current keyword and the preset weight to obtain the third importance.

6. The method according to any one of claims 1 to 5, characterized in that The current indicator sets of the multiple dimensions include a first indicator set, a second indicator set, and a third indicator set; Determining the current product classification result of the product to be analyzed based on the fusion evaluation value includes: Determining, based on the fused evaluation value, a first evaluation value of the product to be analyzed under the first indicator set, a second evaluation value of the product to be analyzed under the second indicator set, and a third evaluation value of the product to be analyzed under the third indicator set; querying, based on the first evaluation value, the second evaluation value, and the third evaluation value, a correspondence between the evaluation values and the product classification results, to obtain a first product classification result of the product to be analyzed under the first indicator set, a second product classification result of the product to be analyzed under the second indicator set, and a third product classification result of the product to be analyzed under the third indicator set; Product categories corresponding to the first product classification result, the second product classification result, and the third product classification result are obtained as the current product classification result.

7. A model-based strategic emerging industry product recommendation device, characterized in that: The device comprises: A set determination module is used to determine a current indicator set of multiple dimensions of a product to be analyzed; the product to be analyzed is used to represent the output of a strategic emerging industry; the current indicator set of each dimension includes multiple current indicators; An importance determination module is used to extract current keywords from the current indicators in the current indicator set of each dimension, and determine the current importance corresponding to the current indicator based on the current keywords; A weight determination module, configured to determine a current weight corresponding to the current indicator according to the current importance; An evaluation value determination module is used to obtain the indicator data of the current indicator, input the indicator data into the trained evaluation value prediction model to obtain the predicted evaluation value corresponding to the indicator data, and fuse the predicted evaluation values according to the current weight to obtain a fused evaluation value corresponding to the current indicator; A result determination module, configured to determine a current product classification result of the product to be analyzed based on the fusion evaluation value; The product recommendation module is used to obtain product demand information of the object to be analyzed, filter out the products to be analyzed whose current product classification results meet the product demand information from each of the products to be analyzed, use them as target products corresponding to the object to be analyzed, and recommend the target products to the object to be analyzed.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. 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 method according to any one of claims 1 to 6 are implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images