Business demand matching method and device, computer device and storage medium

Abstract

The application relates to a service demand matching method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: obtaining current conversation text and each historical conversation text of a user; determining a historical matching score corresponding to each historical conversation text according to a historical hot word matched by the historical conversation text; analyzing the current conversation text to determine a current matching score of the current conversation text; if there is no historical matching score consistent with the current matching score, determining a to-be-fused historical hot word from the historical hot words, fusing the current conversation text and the to-be-fused historical hot word to obtain a fused conversation text; taking the fused conversation text as the current conversation text, returning to the step of analyzing the current conversation text until there is a historical matching score consistent with the current matching score, and matching the service demand of the user according to the historical conversation text corresponding to the consistent historical matching score. The method can improve the efficiency of determining the service demand of the user.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence recognition and classification technology, and in particular to a business demand matching method, apparatus, computer equipment, storage medium and computer program product. Background Technology

[0002] With the development of artificial intelligence technology, business demand matching systems based on user dialogue have emerged. These systems have the capabilities to process dialogue text data and perform semantic analysis. They are suitable for industries such as large-scale knowledge processing, natural language understanding, automatic response, and reasoning. When applied in the field of computer finance, they can identify users' business needs through dialogue. They can not only provide users with the statistical analysis information needed for refined management, but also improve communication efficiency with users, reduce manual call volume, improve user service level and quality, increase operational efficiency, and reduce operating costs.

[0003] Currently, the advent of the information age has led to an exponential increase in dialogue text data. Moreover, as users' business needs increase in number and complexity, traditional business need matching systems can no longer adequately meet users' business needs, resulting in low efficiency in determining users' business needs. Summary of the Invention

[0004] Therefore, it is necessary to provide a business requirement matching method, apparatus, computer equipment, computer-readable storage medium, and computer program product that can improve the efficiency of determining users' business requirements in order to address the above-mentioned technical problems.

[0005] A business requirement matching method, the method comprising:

[0006] Obtain the user's current conversation text and all historical conversation texts;

[0007] Based on the historical hot words matched by each of the historical dialogue texts, the historical matching score corresponding to each of the historical dialogue texts is determined;

[0008] The current dialogue text is analyzed to determine the current matching score of the current dialogue text;

[0009] If there is no historical matching score that matches the current matching score, determine the historical hot words to be fused from the historical hot words, and fuse the current dialogue text with the historical hot words to be fused to obtain the fused dialogue text;

[0010] The fused dialogue text is used as the current dialogue text. The process of analyzing the current dialogue text and determining the current matching score of the current dialogue text continues until a historical matching score that matches the current matching score is found. Based on the historical dialogue text corresponding to the consistent historical matching score, the user's business needs are matched.

[0011] In one embodiment, obtaining the user's current conversation text and each historical conversation text includes:

[0012] Receive a user access request, the user access request carrying user information;

[0013] Based on the user information, retrieve the user's historical conversation texts from a preset database;

[0014] The user's current conversation text is obtained during the interaction.

[0015] In one embodiment, determining the historical matching score corresponding to each historical dialogue text based on the historical hot words matched by each historical dialogue text includes:

[0016] Each of the historical dialogue texts is matched with the user's historical hot word database to obtain the historical hot words matched by each of the historical dialogue texts.

[0017] Based on the historical hot words matched by each of the historical dialogue texts, the historical matching score corresponding to each of the historical dialogue texts is determined.

[0018] In one embodiment, determining the historical matching score corresponding to each historical dialogue text based on the historical hot words matched by each historical dialogue text includes:

[0019] Retrieve the historical hot words matched with each of the stored historical dialogue texts;

[0020] Based on the historical hot words matched by each of the historical dialogue texts, the historical matching score corresponding to each of the historical dialogue texts is determined.

[0021] In one embodiment, determining the historical matching score corresponding to each historical dialogue text based on the historical hot words matched by each historical dialogue text includes:

[0022] Based on the hot word weights of the historical hot words matched by each of the historical dialogue texts, the historical matching score corresponding to each of the historical dialogue texts is determined.

[0023] In one embodiment, the method for determining the historical hot words and their weights includes:

[0024] The historical dialogue texts are segmented into words, and candidate hot words are determined based on the part-of-speech tags of the segmented words.

[0025] Calculate the candidate hot word weight for each of the aforementioned candidate hot words;

[0026] Based on the weights of each candidate hot word, historical hot words are selected from the candidate hot words, and the candidate hot word weights of the determined historical hot words are the hot word weights of the determined historical hot words.

[0027] In one embodiment, calculating the candidate hot word weights of each of the candidate hot words includes:

[0028] Based on the semantic relationships between the candidate hot words, a hot word graph corresponding to the candidate hot words is constructed. The hot word graph includes nodes and edges. The nodes are the candidate hot words, and the edge between two nodes indicates that there is a semantic relationship between the candidate hot words corresponding to the two nodes.

[0029] Based on the aforementioned hot word map, a candidate hot word vector is constructed for each of the candidate hot words;

[0030] Based on the candidate hot word vectors, the candidate hot word weights corresponding to each candidate hot word are calculated iteratively.

[0031] In one embodiment, the step of selecting and determining historical hot words from the candidate hot words according to the weight of each candidate hot word includes: selecting and determining the candidate hot words with a weight greater than a predetermined weight threshold as historical hot words.

[0032] In one embodiment, the step of selecting and determining historical hot words from the candidate hot words according to the weight of each candidate hot word includes: selecting and determining a predetermined number of candidate hot words with the largest weight as historical hot words.

[0033] In one embodiment, the step of selecting and determining historical hot words from the candidate hot words according to the weight of each candidate hot word includes: selecting and determining the candidate hot words with the largest weight and accounting for a predetermined proportion of the candidate hot words as historical hot words.

[0034] In one embodiment, determining the historical matching score corresponding to each historical dialogue text based on the hot word weights of the historical hot words matched by each historical dialogue text includes:

[0035] Based on the hot word weights of the historical hot words matched by each of the historical dialogue texts, and the preset matching coefficients of each of the historical hot words, the historical hot word matching score of each of the historical hot words is determined.

[0036] The sum of the historical hot words matching scores of each of the historical dialogue texts is determined as the historical matching score corresponding to each of the historical dialogue texts.

[0037] In one embodiment, analyzing the current dialogue text to determine the current matching score of the current dialogue text includes:

[0038] Determine the keywords that match the current dialogue text with the preset keyword library, and determine the keyword weight of each keyword;

[0039] Based on the keyword weight of each keyword and the preset matching coefficient of each keyword, the keyword matching score of each keyword is determined;

[0040] Based on the keyword matching scores, the current matching score of the current dialogue text is determined.

[0041] In one embodiment, determining the historical hot words to be merged from the historical hot words includes:

[0042] The historical hot words with the highest weight are selected as the historical hot words to be integrated.

[0043] In one embodiment, after determining that there is no historical matching score consistent with the current matching score, and before determining the historical hot words to be merged from the historical hot words, the method further includes:

[0044] Record the number of times a historical match score that does not match the current match score;

[0045] If the number of matches is less than or equal to the preset number, then proceed to the step of determining the historical hot words to be merged from the historical hot words; otherwise, it is determined that the matching of the user's business needs has failed and the process is switched to manual processing.

[0046] A business demand matching device, the device comprising:

[0047] The dialogue text acquisition module is used to acquire the user's current dialogue text and all historical dialogue texts;

[0048] The historical matching score determination module is used to determine the historical matching score corresponding to each historical dialogue text based on the historical hot words matched by each historical dialogue text.

[0049] The current matching score determination module is used to analyze the current dialogue text and determine the current matching score of the current dialogue text.

[0050] The fusion module is used to determine the historical hot words to be fused from the historical hot words if there is no historical matching score that matches the current matching score, and to fuse the current dialogue text with the historical hot words to be fused to obtain the fused dialogue text.

[0051] The business requirement determination module is used to take the fused dialogue text as the current dialogue text, and if it is determined that there is a historical matching score that is consistent with the current matching score, it matches the user's business requirement according to the historical dialogue text corresponding to the consistent historical matching score.

[0052] In one embodiment, the dialogue text acquisition module includes:

[0053] A request receiving unit is used to receive user access requests, wherein the user access requests carry user information.

[0054] The historical dialogue text acquisition unit is used to acquire the user's historical dialogue text from a preset database based on the user information.

[0055] The current dialogue text acquisition unit is used to acquire the user's current dialogue text during the interaction.

[0056] In one embodiment, the historical matching score determination module includes:

[0057] The first historical hot word matching unit is used to match each of the historical dialogue texts with the user's historical hot word database to obtain the historical hot words matched by each of the historical dialogue texts.

[0058] The historical matching score determination unit is used to determine the historical matching score corresponding to each historical dialogue text based on the historical hot words matched by each historical dialogue text.

[0059] In one embodiment, the historical matching score determination module includes:

[0060] The second historical hot word matching unit is used to obtain the historical hot words matched by each of the stored historical dialogue texts;

[0061] The historical matching score determination unit is used to determine the historical matching score corresponding to each historical dialogue text based on the historical hot words matched by each historical dialogue text.

[0062] In one embodiment, the historical matching score determination unit is further configured to determine the historical matching score corresponding to each historical dialogue text based on the hot word weight of the historical hot words matched by each historical dialogue text.

[0063] In one embodiment, the historical matching score determination unit includes:

[0064] The candidate hot word determination unit is used to segment each of the historical dialogue texts and determine candidate hot words based on the part of speech of the obtained segmented words.

[0065] The candidate hot word weight calculation unit is used to calculate the candidate hot word weight of each of the candidate hot words;

[0066] The historical hot word determination unit is used to filter and determine historical hot words from the candidate hot words according to the weight of each candidate hot word, and the candidate hot word weight of the determined historical hot word is the hot word weight of the determined historical hot word.

[0067] In one embodiment, the candidate hot word weight calculation unit includes:

[0068] A hot word graph construction unit is used to construct a hot word graph corresponding to each candidate hot word based on the semantic relationship between each candidate hot word. The hot word graph includes nodes and edges, where each node is a candidate hot word, and an edge between two nodes indicates that there is a semantic relationship between the candidate hot words corresponding to the two nodes.

[0069] The candidate hot word vector construction unit is used to construct candidate hot word vectors for each of the candidate hot words based on the hot word map;

[0070] The candidate hot word weight determination unit is used to iteratively calculate the candidate hot word weight corresponding to each candidate hot word based on each candidate hot word vector.

[0071] In one embodiment, the candidate hot word weight determination unit is further configured to filter and determine the candidate hot words corresponding to the candidate hot word weights that are greater than a predetermined weight threshold as historical hot words;

[0072] The candidate hot word weight determination unit is further configured to filter and determine the predetermined number of candidate hot words with the largest candidate hot word weight as historical hot words;

[0073] The candidate hot word weight determination unit is further configured to select and determine the candidate hot words with the largest weight and accounting for a predetermined proportion of the candidate hot words as historical hot words.

[0074] In one embodiment, the historical matching score determination unit is further configured to determine the historical hot word matching score of each historical hot word based on the hot word weight of the historical hot words matched by each historical dialogue text and the preset matching coefficient of each historical hot word; and to determine the sum of the historical hot word matching scores of the historical hot words of each historical dialogue text as the historical matching score corresponding to each historical dialogue text.

[0075] In one embodiment, the current matching score determination module includes:

[0076] The keyword weight determination unit is used to determine the keywords that match the current dialogue text with the preset keyword library, and to determine the keyword weight of each keyword.

[0077] The keyword matching score determination unit is used to determine the keyword matching score of each keyword based on the keyword weight of each keyword and the preset matching coefficient of each keyword.

[0078] The current matching score determination unit is used to determine the current matching score of the current dialogue text based on the matching scores of each keyword.

[0079] In one embodiment, the fusion module includes:

[0080] The unit for determining historical hot words to be merged is used to select the historical hot word with the highest weight as the historical hot word to be merged.

[0081] In one embodiment, the device further includes:

[0082] A matching count recording unit is used to record the number of times a historical matching score that does not match the current matching score.

[0083] A matching count determination unit is used to determine whether the matching count is less than or equal to a preset count;

[0084] The matching result determination unit is used to determine that the matching of the user's business needs has failed and to transfer the process to manual processing when the matching number determination unit determines that the matching number is greater than a preset number.

[0085] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the above-described business requirement matching method.

[0086] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described business requirement matching method.

[0087] A computer program product includes a computer program that, when executed by a processor, implements the steps of the above-described business requirement matching method.

[0088] The aforementioned business requirement matching method, apparatus, computer equipment, storage medium, and computer program product acquire the user's current dialogue text and all historical dialogue texts; determine the historical matching score corresponding to each historical dialogue text based on the historical hot words matched by each historical dialogue text; analyze the current dialogue text to determine the current matching score of the current dialogue text; if no historical matching score is found that matches the current matching score, determine the historical hot words to be merged from the historical hot words, merge the current dialogue text with the historical hot words to be merged to obtain the merged dialogue text; use the merged dialogue text as the current dialogue text, return to the step of analyzing the current dialogue text to determine the current matching score of the current dialogue text, until a historical matching score that matches the current matching score is found, and match the user's business requirement based on the historical dialogue text corresponding to the consistent historical matching score. The method described in the above embodiments extracts historical hot words from historical dialogue texts to determine the historical matching scores corresponding to the historical dialogue texts. These scores are then combined with the current matching scores corresponding to the user's current dialogue texts to jointly match the user's business needs. In other words, the user's current business needs are determined based on their frequently used business needs. This effectively improves the accuracy and efficiency of determining the user's business needs. When the accuracy of matching the user's current business needs is high, the efficiency of providing services to the user can also be improved when providing corresponding business services based on those needs. Attached Figure Description

[0089] Figure 1 This is an application environment diagram of the business requirement matching method in one embodiment;

[0090] Figure 2 This is a flowchart illustrating a business requirement matching method in one embodiment;

[0091] Figure 3 This is a schematic diagram illustrating the application of the business requirement matching method in a specific embodiment.

[0092] Figure 4 This is a flowchart illustrating a business requirement matching method in a specific embodiment.

[0093] Figure 5 This is a structural block diagram of a service requirement matching device in one embodiment;

[0094] Figure 6 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0095] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. It should be noted that the acquisition, storage, use, and processing of data in the technical solutions of this application all comply with relevant national laws and regulations.

[0096] In one embodiment, the business demand matching method provided in this application can be applied to both terminal 102 and server 104 in the same environment. Figure 1 As shown in the diagram, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or placed on a cloud or other network server. When a user makes a business inquiry through terminal 102 in text format, including but not limited to voice dialogue and text, server 104 needs to determine the user's business requirements in order to perform corresponding business processing. Specifically, server 104 obtains the user's current dialogue text and all historical dialogue texts; determines the historical matching score corresponding to each historical dialogue text based on the historical hot words matched by each historical dialogue text; analyzes the current dialogue text to determine its current matching score; if no historical matching score matches the current matching score, it determines the historical hot words to be merged from the historical hot words, merges the current dialogue text with the historical hot words to be merged, and obtains the merged dialogue text; it uses the merged dialogue text as the current dialogue text, returns to the step of analyzing the current dialogue text to determine its current matching score, until a historical matching score matches the current matching score, and matches the user's business requirements based on the historical dialogue text corresponding to the consistent historical matching score.

[0097] In one embodiment, the business requirement matching method provided in this application only involves terminal 102 or server 104 in the application environment. Specifically, the user can directly conduct business inquiries in text form on terminal 102 or server 104. Terminal 102 or server 104 obtains the user's current dialogue text and each historical dialogue text; determines the historical matching score corresponding to each historical dialogue text based on the historical hot words matched by each historical dialogue text; analyzes the current dialogue text to determine the current matching score of the current dialogue text; if there is no historical matching score consistent with the current matching score, determines the historical hot words to be merged from the historical hot words, merges the current dialogue text with the historical hot words to be merged to obtain the merged dialogue text; uses the merged dialogue text as the current dialogue text, and returns to the step of analyzing the current dialogue text to determine the current matching score of the current dialogue text, until a historical matching score consistent with the current matching score is found, and matches the user's business requirement based on the historical dialogue text corresponding to the consistent historical matching score.

[0098] The terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. The server 104 can be implemented using a standalone server or a server cluster consisting of multiple servers.

[0099] In one embodiment, such as Figure 2 As shown, a business requirement matching method is provided, which is applied to... Figure 1 Taking terminal 102 and / or server 104 as examples, the following is an explanation:

[0100] Step S202: Obtain the user's current dialogue text and all historical dialogue texts.

[0101] In one embodiment, a user can conduct business inquiries in text form, which includes, but is not limited to, voice conversations and text messages. The text currently used for the user's business inquiry is referred to as the current conversation text. For example, when a user conducts a business inquiry via voice conversation, they can make a phone call. The party providing the business inquiry will receive the call and, after the user has spoken the conversation, will be able to obtain the current conversation text. When a user conducts a business inquiry via text messages, they can send text related to the business inquiry using a social media account. The party providing the business inquiry will receive the text sent by the user. To improve the accuracy of determining the user's business needs, the party providing the business inquiry needs to combine the user's current conversation text with historical conversation texts to determine the user's business needs. The party providing business inquiry in this embodiment includes, but is not limited to, financial institutions.

[0102] In one embodiment, the text in which a user has already made a business inquiry is referred to as historical dialogue text. This means that the user's business needs corresponding to the historical dialogue text are already determined, and the user's historical dialogue text includes at least one such text. The user's historical dialogue text can be a single sentence or a paragraph composed of multiple single sentences; this is not limited here. The user's historical dialogue text and the business needs matched with it are stored in a preset database. Specifically, a user access request is received, carrying user information. Based on the user information, the user's historical dialogue texts are retrieved from the preset database, and the user's current dialogue text is retrieved during the interaction. The user access request can be an incoming call, and the user information carried in the request can be the user's phone number. The user access request can also be a text message sent by the user, and the user information carried in the request can be the user's social media account information.

[0103] Step S204: Determine the historical matching score for each historical dialogue text based on the historical hot words matched by each historical dialogue text.

[0104] In one embodiment, "hot words" refer to popular terms that represent issues and things that users are generally concerned about within a certain period, and their main forms of expression include language and text. Hot words in a user's historical dialogue text are called historical hot words. Historical dialogue texts are matched with historical hot words. To determine the relevance between current and historical dialogue texts and improve the accuracy of identifying user business needs, historical dialogue texts are quantified, i.e., historical matching scores are determined for each historical dialogue text. Specifically, based on the historical hot words matched by each historical dialogue text, a historical matching score is determined for each historical dialogue text.

[0105] In one embodiment, the historical matching score corresponding to the historical dialogue text needs to be determined based on the historical hot words in the historical dialogue text. Specifically, historical hot words corresponding to the user can be pre-stored to form a user's historical hot word library. Based on the user's historical hot word library, after obtaining the historical dialogue text, the historical hot words matching each historical dialogue text are extracted and determined. Specifically, each historical dialogue text is matched with the user's historical hot word library to obtain the historical hot words matching each historical dialogue text. Based on the historical hot words matching each historical dialogue text, the historical matching score corresponding to each historical dialogue text is determined. Alternatively, the determined historical dialogue texts and their matching historical hot words can be stored together. After obtaining the historical dialogue text, the stored historical hot words can be directly determined. Specifically, the stored historical hot words matching each historical dialogue text are obtained, and based on the historical hot words matching each historical dialogue text, the historical matching score corresponding to each historical dialogue text is determined.

[0106] In one embodiment, the historical matching score corresponding to the historical dialogue text can be determined based on the hot word weights of historical hot words. The hot word weights of historical hot words can be understood as the importance of the historical hot words within the historical dialogue statements. Specifically, the historical matching score corresponding to each historical dialogue text is determined based on the hot word weights of the historical hot words matched in each historical dialogue text.

[0107] In one embodiment, the historical matching score for each historical dialogue text is determined based on the hot word weights of the historical hot words matched in each historical dialogue text. Specifically, the historical hot word matching score for each historical hot word is determined based on its hot word weight and a preset matching coefficient. The preset matching coefficient for each historical hot word can be determined by the party providing business consultation, depending on the actual usage scenario, and is generally set between 0 and 1, specifically 0.5. The preset matching coefficients for each historical hot word can be the same or different. Then, the sum of the historical hot word matching scores for each historical dialogue text is determined as the historical matching score for each historical dialogue text. For example, if a historical dialogue text contains historical hot word 1 and historical hot word 2, then the matching score for historical hot word 1 is the product of its hot word weight and the preset matching coefficient, and the matching score for historical hot word 2 is the product of its hot word weight and the preset matching coefficient. The historical matching score for the historical dialogue text is the sum of the matching scores for historical hot word 1 and historical hot word 2.

[0108] In one embodiment, historical hot words can be extracted using a keyword extraction algorithm, namely the TextRank algorithm. The TextRank algorithm constructs a network based on the adjacency relationships between words, iteratively calculates the weight of each node in the network graph, and ranks the weights to obtain historical hot words. Specifically, the determination of historical hot words and their weights includes:

[0109] Step S302: Segment each historical dialogue text into words, and determine candidate hot words based on the part of speech of the obtained words.

[0110] In one embodiment, if the user's historical dialogue text consists of single-sentence historical dialogue statements, then the historical dialogue statements can be directly segmented into words. If the user's historical dialogue text consists of a historical dialogue paragraph composed of multiple single-sentence historical dialogue statements, then the historical dialogue paragraph needs to be segmented into single-sentence historical dialogue statements, and then the historical dialogue statements need to be segmented into words. After segmenting each historical dialogue text to obtain each word, the words with preset parts of speech such as nouns, adjectives, and verbs in each word segment are determined as candidate hot words.

[0111] Step S304: Calculate the candidate hot word weight for each candidate hot word.

[0112] In one embodiment, a network is constructed based on the adjacency relationships between candidate hot words. This constructed network is called a hot word graph, and the nodes in the hot word graph are candidate hot words. By iteratively calculating the weight of each node in the hot word graph, the candidate hot word weight can be obtained. Specifically, based on the semantic relationships between candidate hot words, a hot word graph corresponding to the candidate hot words is constructed. The hot word graph includes nodes and edges. Nodes are candidate hot words, and an edge between two nodes indicates that there is a semantic relationship between the candidate hot words corresponding to the two nodes. Based on the hot word graph, a candidate hot word vector is constructed for each candidate hot word. Here, an edge between any two nodes is constructed based on the co-occurrence relationship to form a candidate hot word vector. And based on each candidate hot word vector, the candidate hot word weight corresponding to each candidate hot word is iteratively calculated. The formula for iterative calculation is expressed as:

[0113]

[0114] Among them, WS(V i ) represents the node V where the candidate hot words are located. i Candidate hot word weights, In(V i ) represents the node V where the candidate hot words are located. i The set of predecessor nodes, Out(V) j ) represents the node V where the candidate hot words are located. i The set of successor nodes, d represents the damping coefficient used for smoothing, w ji For node V i and node V j The weight of the edges between them.

[0115] Step S306: Based on the weight of each candidate hot word, select and determine the historical hot words from the candidate hot words. The candidate hot word weight of the determined historical hot words is the hot word weight of the determined historical hot words.

[0116] In one embodiment, historical hot words are selected from the candidate hot words based on their weights. Specifically, candidate hot words with larger weights are identified as historical hot words. Specifically, candidate hot words with weights greater than a predetermined weight threshold are selected as historical hot words. That is, candidate hot words are sorted according to their weights, and those with weights greater than the predetermined weight threshold are selected as historical hot words. Alternatively, a predetermined number of candidate hot words with the largest weights can be selected as historical hot words. That is, candidate hot words are sorted according to their weights; if sorted from smallest to largest, a predetermined number of candidate hot words at the end are selected as historical hot words; if sorted from largest to smallest, a predetermined number of candidate hot words at the beginning are selected as historical hot words. Finally, candidate hot words with the largest weights and representing a predetermined proportion of the candidate hot words can also be selected as historical hot words. That is, to identify candidate hot words with a relatively large weight according to a predetermined proportion and to designate them as historical hot words.

[0117] Step S206: Analyze the current dialogue text and determine the current matching score of the current dialogue text.

[0118] In one embodiment, to determine the relevance between the current dialogue text and historical dialogue texts and improve the accuracy of identifying user business needs, it is also necessary to quantify the current dialogue text, that is, to determine the current matching score corresponding to the current dialogue text. Specifically, the current dialogue text is analyzed to determine the current matching score of the current dialogue text.

[0119] In one embodiment, the current matching score of the current dialogue text can be determined in a manner similar to that used for historical dialogue texts. Specifically, the current matching score is determined by identifying each keyword in the current dialogue text. A pre-stored database of frequently used keywords is used to form a pre-defined keyword library. Specifically, the keywords that match the current dialogue text with the pre-defined keyword library are determined, and the keyword weight of each keyword is determined. Based on the keyword weight and the pre-defined matching coefficient of each keyword, the keyword matching score of each keyword is determined. Based on the keyword matching scores, the current matching score of the current dialogue text is determined. The pre-defined matching coefficient of each keyword can be determined by the party providing business consultation, depending on the actual usage scenario. It is generally set between 0 and 1, specifically 0.5. The pre-defined matching coefficients of each keyword can be the same or different. For example, if a current dialogue text contains keyword 1 and keyword 2, then the matching score of keyword 1 is the product of its keyword weight and the pre-defined matching coefficient, and the matching score of keyword 2 is the product of its keyword weight and the pre-defined matching coefficient. The current matching score of the current dialogue text is the sum of the matching scores of keyword 1 and keyword 2.

[0120] In one embodiment, the current dialogue text can also be analyzed to directly determine the current matching score, i.e., the current matching score can be determined by performing semantic analysis on the current dialogue text. Semantic analysis can involve performing syntactic, pragmatic, and contextual analysis on the current dialogue statements to comprehensively determine the current matching score.

[0121] Step S208: If there is no historical matching score that matches the current matching score, determine the historical hot words to be fused from the historical hot words, and fuse the current dialogue text with the historical hot words to be fused to obtain the fused dialogue text.

[0122] In one embodiment, after determining the current matching score of the current dialogue text and the historical matching scores of each historical dialogue text, the current matching score is compared with each historical matching score to determine whether there is a historical matching score that matches the current matching score. Here, "matching" does not mean that the current matching score and the historical matching score are exactly the same; the error between the two is sufficient if it is less than a preset error threshold. The preset error threshold can be set to 10%. Specifically, the error between the current matching score and each historical matching score is calculated to determine whether there is a historical matching score with an error less than the preset error threshold. If there is a historical matching score with an error less than the preset error threshold, it is determined that there is a historical matching score that matches the current matching score. If there are two or more historical matching scores that match the current matching score, the historical matching score with the smallest error is determined as the matching historical matching score.

[0123] In one embodiment, if no historical matching score exists with an error less than a preset error threshold, the current dialogue text needs to be fused. Specifically, a historical hot word to be fused is determined from historical hot words. There is only one historical hot word to be fused, and the current dialogue text is fused with the historical hot word to be fused. Among them, the historical hot word with the highest weight is selected as the historical hot word to be fused. After determining the historical hot word to be fused, it is fused with the current dialogue text to obtain the fused dialogue text.

[0124] Step S210: Take the merged dialogue text as the current dialogue text, return to the step of analyzing the current dialogue text and determining the current matching score of the current dialogue text, until there is a historical matching score that matches the current matching score, and match the user's business needs according to the historical dialogue text corresponding to the consistent historical matching score.

[0125] In one embodiment, the merged dialogue text is used as the current dialogue text, and the process returns to step S206, which involves analyzing the current dialogue text, determining its current matching score, and determining a new current matching score for the current dialogue text, until a historical matching score matching the current score is found. Since the business needs of the user corresponding to the historical dialogue text are already determined, once a historical matching score matching the current score is determined, the historical dialogue text matching the current dialogue text is identified, thus determining the user's business needs. Specifically, if a historical matching score matching the current score exists, the user's business needs are matched based on the historical dialogue text corresponding to the consistent historical matching score.

[0126] In one embodiment, to avoid endlessly fusing the current dialogue text and improve the efficiency of matching user business needs, after determining that there is no historical matching score consistent with the current matching score, and before determining the historical hot words to be fused from the historical hot words, the method further includes: recording the number of times there is no historical matching score consistent with the current matching score; if the number of matching is less than or equal to a preset number, then proceeding to the step of determining the historical hot words to be fused from the historical hot words; otherwise, determining the user's business needs fails and is switched to manual processing, i.e., manually determining the user's business needs. The preset number of times can be determined according to the actual usage scenario, specifically set to 5 to 10 times.

[0127] In one embodiment, manual processing includes multiple manual processing units, each corresponding to a different business requirement. When matching a user's business requirement fails and the process is transferred to manual processing, the historical dialogue text with the highest degree of consistency with the current dialogue text can be identified and transferred to the manual processing unit corresponding to the business requirement of that historical dialogue text, thereby providing precise and intelligent manual services. Specifically, the historical dialogue text with the lowest error between its current matching score and the current dialogue text can be identified as the historical dialogue text with the highest degree of consistency with the current dialogue text.

[0128] In the aforementioned business requirement matching method, the following steps are taken: First, the user's current dialogue text and all historical dialogue texts are obtained. Then, based on the historical hot words matched by each historical dialogue text, a historical matching score is determined for each historical dialogue text. Next, the current dialogue text is analyzed to determine its current matching score. If no historical matching score matches the current score, a historical hot word to be merged is identified from the historical hot words. The current dialogue text is then merged with the historical hot word to obtain a merged dialogue text. This merged dialogue text is used as the current dialogue text, and the process of analyzing the current dialogue text and determining its current matching score is repeated until a historical matching score matches the current score. Finally, the user's business requirement is matched based on the historical dialogue text corresponding to the consistent historical matching score. By extracting historical hot words from historical dialogue texts, determining the historical matching score corresponding to each historical dialogue text, and combining this with the current matching score of the user's current dialogue text to match the user's business requirement, the accuracy and efficiency of determining the user's business requirement can be effectively improved, thereby enhancing service efficiency.

[0129] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description, in conjunction with the accompanying drawings and a specific embodiment, will further illustrate this application. It should be understood that the specific embodiment described herein is merely illustrative and not intended to limit the scope of this application.

[0130] In one specific embodiment, such as Figure 3 The diagram illustrates the application of a business demand matching method, primarily involving users and financial institutions. Users use their terminals to call the financial institution's customer service hotline for relevant business inquiries. Specifically, for example... Figure 4 The diagram below illustrates the process of matching business requirements. The steps are as follows:

[0131] Receive incoming calls from users, obtain phone number information from the incoming calls, retrieve the user's historical conversation texts from a preset database based on the user's phone number information, and obtain the user's current conversation text during the interaction process;

[0132] Each historical dialogue text is segmented into words, and words with preset parts of speech such as nouns, adjectives, and verbs are retained as candidate hot words. Based on the semantic relationship between each candidate hot word, a hot word graph corresponding to the candidate hot words is constructed. The hot word graph includes nodes and edges. Nodes are candidate hot words, and the edge between two nodes indicates that there is a semantic relationship between the candidate hot words corresponding to the two nodes.

[0133] Based on the hot word graph, edges are constructed between any two nodes according to co-occurrence relationships, forming candidate hot word vectors for each candidate hot word. Based on each candidate hot word vector, the candidate hot word weights corresponding to each candidate hot word are iteratively calculated. The formula for the iterative calculation is expressed as:

[0134]

[0135] Among them, WS(V i ) represents the node V where the candidate hot words are located. i Candidate hot word weights, In(V i ) represents the node V where the candidate hot words are located. i The set of predecessor nodes, Out(V) j ) represents the node V where the candidate hot words are located. i The set of successor nodes, d represents the damping coefficient used for smoothing, w ji For node V i and node V j The weight of the edges between them;

[0136] Based on the weight of each candidate hot word, they are arranged in descending order, and a predetermined number of candidate hot words at the top are selected as historical hot words. The candidate hot word weights of the determined historical hot words are the hot word weights of the determined historical hot words.

[0137] Based on the hot word weights of the historical hot words matched in each historical dialogue text, the historical matching score corresponding to each historical dialogue text is determined. Wherein, if the historical dialogue text contains historical hot word 1 and historical hot word 2, then the matching score of historical hot word 1 is the product of the hot word weight of historical hot word 1 and the preset matching coefficient, the matching score of historical hot word 2 is the product of the hot word weight of historical hot word 2 and the preset matching coefficient, and the historical matching score corresponding to the historical dialogue text is the sum of the matching scores of historical hot word 1 and historical hot word 2.

[0138] Determine the keywords that match the current dialogue text with the preset keyword library, and determine the keyword weight of each keyword; determine the keyword matching score of each keyword based on the keyword weight and the preset matching coefficient of each keyword; determine the current matching score of the current dialogue text based on the keyword matching scores of each keyword, where the current dialogue text contains keyword 1 and keyword 2, then the matching score of keyword 1 is the product of the keyword weight of keyword 1 and the preset matching coefficient, the matching score of keyword 2 is the product of the keyword weight of keyword 2 and the preset matching coefficient, and the current matching score corresponding to the current dialogue text is the sum of the matching scores of keyword 1 and keyword 2.

[0139] Calculate the error between the current matching score and each historical matching score, and determine whether there are any historical matching scores with an error less than a preset error threshold, where the preset error threshold is set to 10%;

[0140] If there is no historical match score that matches the current match score, record the number of matches for which there is no historical match score that matches the current match score;

[0141] If the number of matching attempts exceeds the preset number, the user's business request will be determined to have failed and will be transferred to manual processing. The preset number of matching attempts is set to 5.

[0142] If the number of matches is less than or equal to the preset number, the historical hot word with the highest weight will be used as the historical hot word to be merged. The current dialogue text will be merged with the historical hot word to be merged to obtain the merged dialogue text.

[0143] The process involves taking the merged dialogue text as the current dialogue text, analyzing the current dialogue text, determining the current matching score, and continuing until a historical matching score is found that matches the current matching score. Based on the historical dialogue text corresponding to the consistent historical matching score, the user's business needs are matched.

[0144] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0145] Based on the same inventive concept, this application also provides a business requirement matching apparatus for implementing the business requirement matching method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more embodiments of the business requirement matching apparatus provided below can be found in the limitations of the business requirement matching method described above, and will not be repeated here.

[0146] In one embodiment, such as Figure 5 As shown, a business requirement matching device is provided, including: a dialogue text acquisition module 510, a historical matching score determination module 520, a current matching score determination module 530, a fusion module 540, and a business requirement determination module 550, wherein:

[0147] The dialogue text acquisition module 510 is used to acquire the user's current dialogue text and all historical dialogue texts.

[0148] The historical matching score determination module 520 is used to determine the historical matching score corresponding to each historical dialogue text based on the historical hot words matched by each historical dialogue text.

[0149] The current matching score determination module 530 is used to analyze the current dialogue text and determine the current matching score of the current dialogue text.

[0150] The fusion module 540 is used to determine the historical hot words to be fused from the historical hot words if there is no historical matching score that matches the current matching score, and to fuse the current dialogue text with the historical hot words to be fused to obtain the fused dialogue text.

[0151] The business requirement determination module 550 is used to take the fused dialogue text as the current dialogue text, and if it is determined that there is a historical matching score that is consistent with the current matching score, it matches the user's business requirement according to the historical dialogue text corresponding to the consistent historical matching score.

[0152] In one embodiment, the dialogue text acquisition module 510 includes:

[0153] The request receiving unit is used to receive user access requests, which carry user information.

[0154] The historical dialogue text acquisition unit is used to acquire the user's historical dialogue text from a preset database based on the user information.

[0155] The current dialogue text acquisition unit is used to acquire the user's current dialogue text during the interaction.

[0156] In one embodiment, the historical matching score determination module 520 includes:

[0157] The first historical hot word matching unit is used to match each of the historical dialogue texts with the user's historical hot word database to obtain the historical hot words matched by each of the historical dialogue texts.

[0158] The historical matching score determination unit is used to determine the historical matching score corresponding to each historical dialogue text based on the historical hot words matched by each historical dialogue text.

[0159] In one embodiment, the historical matching score determination module 520 includes:

[0160] The second historical hot word matching unit is used to obtain the historical hot words matched by each of the stored historical dialogue texts.

[0161] The historical matching score determination unit is used to determine the historical matching score corresponding to each historical dialogue text based on the historical hot words matched by each historical dialogue text.

[0162] In one embodiment, the historical matching score determination unit is further configured to determine the historical matching score corresponding to each historical dialogue text based on the hot word weight of the historical hot words matched by each historical dialogue text.

[0163] In one embodiment, the historical matching score determination unit includes:

[0164] The candidate hot word determination unit is used to segment each of the historical dialogue texts and determine candidate hot words based on the part-of-speech of the obtained segmented words.

[0165] The candidate hot word weight calculation unit is used to calculate the candidate hot word weight of each of the candidate hot words.

[0166] The historical hot word determination unit is used to filter and determine historical hot words from the candidate hot words according to the weight of each candidate hot word, and the candidate hot word weight of the determined historical hot word is the hot word weight of the determined historical hot word.

[0167] In one embodiment, the candidate hot word weight calculation unit includes:

[0168] A hot word graph construction unit is used to construct a hot word graph corresponding to each candidate hot word based on the semantic relationship between each candidate hot word. The hot word graph includes nodes and edges, where each node is a candidate hot word, and an edge between two nodes indicates that there is a semantic relationship between the candidate hot words corresponding to the two nodes.

[0169] The candidate hot word vector construction unit is used to construct candidate hot word vectors for each of the candidate hot words based on the hot word map.

[0170] The candidate hot word weight determination unit is used to iteratively calculate the candidate hot word weight corresponding to each candidate hot word based on each candidate hot word vector.

[0171] In one embodiment, the candidate hot word weight determination unit is further configured to filter and determine the candidate hot words corresponding to the candidate hot word weights that are greater than a predetermined weight threshold as historical hot words.

[0172] In one embodiment, the candidate hot word weight determination unit is further configured to filter and determine the predetermined number of candidate hot words with the largest candidate hot word weight as historical hot words.

[0173] In one embodiment, the candidate hot word weight determination unit is further configured to filter and determine the candidate hot words with the largest weight and accounting for a predetermined proportion of the candidate hot words as historical hot words.

[0174] In one embodiment, the historical matching score determination unit is further configured to determine the historical hot word matching score of each historical hot word based on the hot word weight of the historical hot words matched by each historical dialogue text and the preset matching coefficient of each historical hot word; and to determine the sum of the historical hot word matching scores of the historical hot words of each historical dialogue text as the historical matching score corresponding to each historical dialogue text.

[0175] In one embodiment, the current matching score determination module 530 includes:

[0176] The keyword weight determination unit is used to determine the keywords that match the current dialogue text with the preset keyword library, and to determine the keyword weight of each keyword.

[0177] The keyword matching score determination unit is used to determine the keyword matching score of each keyword based on the keyword weight of each keyword and the preset matching coefficient of each keyword.

[0178] The current matching score determination unit is used to determine the current matching score of the current dialogue text based on the matching scores of each keyword.

[0179] In one embodiment, the fusion module 540 includes:

[0180] The unit for determining historical hot words to be merged is used to select the historical hot word with the highest weight as the historical hot word to be merged.

[0181] In one embodiment, the business demand matching device further includes:

[0182] The matching count recording unit is used to record the number of times that there is no historical matching score that matches the current matching score.

[0183] The matching count determination unit is used to determine whether the matching count is less than or equal to a preset count.

[0184] The matching result determination unit is used to determine that the matching of the user's business needs has failed and to transfer the process to manual processing when the matching number determination unit determines that the matching number is greater than a preset number.

[0185] Each module in the aforementioned business demand matching device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0186] In one embodiment, a computer device is provided, the internal structure of which can be shown in the figure below. Figure 6 As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a business requirement matching method. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0187] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0188] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described business requirement matching method.

[0189] In one embodiment, a computer-readable storage medium is provided that stores a computer program thereon, which, when executed by a processor, implements the steps of the above-described business requirement matching method.

[0190] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the business requirement matching method described above.

[0191] 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, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0192] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, 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 many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0193] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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.

[0194] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method of matching business requirements, characterized by, The method comprises: obtaining current dialogue text of a user and each historical dialogue text, and a business requirement of the user corresponding to the historical dialogue text being determined; determining a historical matching score corresponding to each historical dialogue text according to a historical hot word matched by each historical dialogue text, the historical hot word being obtained by matching each historical dialogue text with a historical hot word library of the user; the historical matching score corresponding to the historical dialogue text being a sum of matching scores of each historical hot word in the historical dialogue text, and the matching score of the historical hot word being a product of a hot word weight and a preset matching coefficient; analyzing the current dialogue text to determine a current matching score of the current dialogue text; the current matching score of the current dialogue text being a sum of matching scores of each keyword in the current dialogue text, and the matching score of the keyword being a product of a keyword weight and a preset matching coefficient; if there is no historical matching score consistent with the current matching score, determining a to-be-fused historical hot word from the historical hot word, fusing the current dialogue text and the to-be-fused historical hot word to obtain a fused dialogue text; taking the fused dialogue text as the current dialogue text, returning to the step of analyzing the current dialogue text to determine the current matching score of the current dialogue text, until there is a historical matching score consistent with the current matching score, and matching the business requirement of the user according to the historical dialogue text corresponding to the consistent historical matching score; wherein the determination manner of the historical hot word and the hot word weight of the historical hot word comprises: performing word segmentation on each historical dialogue text, determining a candidate hot word according to the part of speech of the obtained word segmentation, constructing a hot word graph corresponding to the candidate hot word according to the semantic relationship between each candidate hot word, constructing an edge between any two nodes in the hot word graph according to the co-occurrence relationship in the hot word graph to form a candidate hot word vector, iteratively calculating a candidate hot word weight corresponding to each candidate hot word according to each candidate hot word vector, and screening and determining a historical hot word from the candidate hot word according to each candidate hot word weight.

2. The business requirement matching method of claim 1, wherein, The obtaining of the current dialogue text of the user and each historical dialogue text comprises: receiving a user access request, the user access request carrying user information; obtaining each historical dialogue text of the user in a preset database according to the user information; obtaining the current dialogue text of the user in an interactive process.

3. The method of claim 1, wherein, The determination of the historical matching score corresponding to each historical dialogue text according to the historical hot word matched by each historical dialogue text comprises: matching each historical dialogue text with the historical hot word library of the user respectively to obtain the historical hot word matched by each historical dialogue text respectively; determining the historical matching score corresponding to each historical dialogue text according to the historical hot word matched by each historical dialogue text respectively.

4. The method of claim 1, wherein, The determination of the historical matching score corresponding to each historical dialogue text according to the historical hot word matched by each historical dialogue text comprises: obtaining the historical hot word matched by each historical dialogue text stored; According to the historical hot words matched by each of the historical dialogue texts, historical matching scores corresponding to each of the historical dialogue texts are determined.

5. The business demand matching method according to claim 3 or 4, characterized by, The historical matching scores corresponding to each of the historical dialogue texts are determined according to the historical hot words matched by each of the historical dialogue texts, including: The historical matching scores corresponding to each of the historical dialogue texts are determined according to the historical hot words matched by each of the historical dialogue texts.

6. The method of claim 1, wherein, The historical matching scores corresponding to each of the historical dialogue texts are determined according to the historical hot words matched by each of the historical dialogue texts.

7. The method of claim 1, wherein, The hot word graph includes nodes and edges, the nodes are the candidate hot words, and the edges of two nodes represent that there is a semantic relationship between the candidate hot words corresponding to the two nodes. The historical hot words are determined from the candidate hot words according to the candidate hot word weights, including any one of the following: The first item: the candidate hot words corresponding to the candidate hot word weights greater than a predetermined weight threshold are determined as historical hot words; The second item: a predetermined number of candidate hot words with the largest candidate hot word weights are determined as historical hot words; 8. The method of claim 5, wherein, The third item: the candidate hot words with the largest candidate hot word weights and occupying a predetermined proportion of the candidate hot words are determined as historical hot words. The historical matching scores corresponding to each of the historical dialogue texts are determined according to the historical hot words matched by each of the historical dialogue texts, including: The historical hot word matching scores of each of the historical hot words are determined according to the historical hot word weights of the historical hot words matched by each of the historical dialogue texts and a preset matching coefficient of each of the historical hot words; 9. The method of claim 1, wherein, The sum of the historical hot word matching scores of the historical hot words of each of the historical dialogue texts is determined as the historical matching score corresponding to each of the historical dialogue texts. The current matching score of the current dialogue text is determined by analyzing the current dialogue text, including: Determine the keywords matched by the current dialogue text and a preset keyword library, and determine the keyword weights of each keyword; Determine the keyword matching scores of each keyword according to the keyword weights of each keyword and a preset matching coefficient of each keyword; 10. The method of claim 1, wherein, Determine the current matching score of the current dialogue text based on each keyword matching score. The to-be-fused historical hot word is determined from the historical hot words, including:

11. The method of claim 1, wherein, The historical hot word with the largest historical hot word weight is taken as the to-be-fused historical hot word. After it is determined that there is no historical matching score consistent with the current matching score, before the to-be-fused historical hot word is determined from the historical hot words, it further includes: Record the matching number of the historical matching score that is not consistent with the current matching score; 12. A service requirement matching apparatus characterized by comprising: If the matching number is less than or equal to a preset number, enter the step of determining the to-be-fused historical hot word from the historical hot words, otherwise, determine that the user's business demand fails to be matched, and transfer to manual processing. The device includes: A dialogue text acquisition module is configured to acquire a current dialogue text of a user and each historical dialogue text, and a business demand of a user corresponding to a historical dialogue text has been determined; The historical matching score determination module is configured to determine a historical matching score corresponding to each historical dialogue text according to historical hot words matched by each historical dialogue text, the historical hot words being obtained by matching each historical dialogue text with a historical hot word library of the user; the historical matching score corresponding to the historical dialogue text being a sum of matching scores of the historical hot words in the historical dialogue text, the matching score of the historical hot word being a product of a hot word weight and a preset matching coefficient; The current matching score determination module is configured to analyze the current dialogue text to determine a current matching score of the current dialogue text; the current matching score of the current dialogue text being a sum of matching scores of keywords in the current dialogue text, the matching score of the keyword being a product of a keyword weight and a preset matching coefficient; The fusion module is configured to determine a to-be-fused historical hot word from the historical hot words if there is no historical matching score consistent with the current matching score, and fuse the current dialogue text and the to-be-fused historical hot word to obtain a fused dialogue text. The business requirement determination module is configured to take the fused dialogue text as the current dialogue text, and match a business requirement of the user according to the historical dialogue text corresponding to the consistent historical matching score if it is determined that there is the historical matching score consistent with the current matching score. The determination manner of the historical hot words and the hot word weight of the historical hot word includes: performing word segmentation on each historical dialogue text, determining candidate hot words according to the part of speech of the segmented words, constructing a hot word graph corresponding to each candidate hot word according to the semantic relationship between the candidate hot words, constructing an edge between any two nodes in the hot word graph according to the co-occurrence relationship in the hot word graph to form a candidate hot word vector, iteratively calculating a candidate hot word weight corresponding to each candidate hot word according to each candidate hot word vector, and screening and determining historical hot words from the candidate hot words according to each candidate hot word weight.

13. The service requirement matching apparatus according to claim 12, wherein The dialogue text acquisition module includes: A request receiving unit configured to receive a user access request, the user access request carrying user information; A historical dialogue text acquisition unit configured to acquire historical dialogue texts of the user from a preset database according to the user information; A current dialogue text acquisition unit configured to acquire a current dialogue text of the user in an interaction process.

14. The service requirement matching apparatus according to claim 12, wherein The historical matching score determination module includes: A historical hot word first matching unit configured to match each historical dialogue text with the historical hot word library of the user respectively to obtain historical hot words matched by each historical dialogue text; A historical matching score determination unit configured to determine a historical matching score corresponding to each historical dialogue text according to the historical hot words matched by each historical dialogue text.

15. The service requirement matching apparatus according to claim 12, wherein The historical matching score determination module includes: A historical hot word second matching unit configured to acquire historical hot words matched by each historical dialogue text stored; A historical matching score determination unit configured to determine a historical matching score corresponding to each historical dialogue text according to the historical hot words matched by each historical dialogue text. 16.The service demand matching apparatus of claim 14 or 15, wherein: The historical matching score determination unit is further configured to determine a historical matching score corresponding to each of the historical dialogue texts according to a hotword weight of a historical hotword matched by the historical dialogue text.

17. The business demand matching apparatus of claim 12, wherein, The hotword graph comprises nodes and edges, the nodes being the candidate hotwords, and an edge between two nodes representing that there is a semantic relationship between the candidate hotwords corresponding to the two nodes. 18.The service demand matching apparatus of claim 12, wherein: The candidate hotword weight determination unit is further configured to determine, as a historical hotword, a candidate hotword corresponding to a candidate hotword weight greater than a predetermined weight threshold. The candidate hotword weight determination unit is further configured to determine, as a historical hotword, a predetermined number of candidate hotwords with the greatest candidate hotword weights. The candidate hotword weight determination unit is further configured to determine, as a historical hotword, a candidate hotword with the greatest candidate hotword weight and occupying a predetermined proportion of the candidate hotwords. 19.The service demand matching apparatus of claim 16, wherein: The historical matching score determination unit is further configured to determine a historical hotword matching score of each historical hotword according to a hotword weight of the historical hotword matched by each historical dialogue text and a preset matching coefficient of the historical hotword. The historical matching score determination unit is further configured to determine a sum of the historical hotword matching scores of the historical hotwords of each historical dialogue text as a historical matching score corresponding to the historical dialogue text.

20. The business demand matching apparatus of claim 12, wherein, The current matching score determination module comprises: A keyword weight determination unit configured to determine keywords matched by the current dialogue text with a preset keyword library and determine keyword weights of the keywords. A keyword matching score determination unit configured to determine keyword matching scores of the keywords according to the keyword weights of the keywords and preset matching coefficients of the keywords. A current matching score determination unit configured to determine a current matching score of the current dialogue text based on the keyword matching scores.

21. The business demand matching apparatus of claim 12, wherein, The fusion module comprises: A to-be-fused historical hotword determination unit configured to determine, as a to-be-fused historical hotword, a historical hotword with the greatest historical hotword weight.

22. The business demand matching apparatus of claim 12, wherein, The apparatus further comprises: A matching frequency recording unit configured to record a matching frequency of the historical matching scores that are not consistent with the current matching score. A matching frequency determination unit configured to determine whether the matching frequency is less than or equal to a preset frequency. A matching result determination unit configured to determine that matching the service demand of the user fails and switch to manual processing when the matching frequency determination unit determines that the matching frequency is greater than the preset frequency. 23.A computer device, comprising a memory and a processor, wherein the memory stores a computer program. The processor implements the steps of the service demand matching method of any one of claims 1 to 11 when executing the computer program.

24. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program implements the steps of the service demand matching method of any one of claims 1 to 11 when executed by the processor.

25. A computer program product comprising a computer program, characterised in that, The computer program implements the steps of the service demand matching method of any one of claims 1 to 11 when executed by the processor.

Images