Project intelligent consultation system and method based on artificial intelligence

Abstract

The invention relates to the technical field of project recommendation, in particular to an intelligent project consultation system and method based on artificial intelligence, and the system comprises an information collection unit, a consultation matching unit, a project scoring unit, a feature extraction unit, a project classification unit, a project recommendation unit and a score updating unit. Specifically, the information acquisition unit is used for acquiring project consultation information of a target consultation user based on a preset project consultation template, and transmitting the project consultation information of the target consultation user to the consultation matching unit. Specific consultation requirements of a user are acquired through the information acquisition unit, matching with a preset project library is performed through the consultation matching unit, the system can provide a candidate project set most relevant to the user according to the requirements of the user, and then, through historical score information and score feature extraction of the project scoring unit and the feature extraction unit, the candidate project set can be obtained. Therefore, the recommended items are more personalized and accurate, and the user experience is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of project recommendation, and in particular to an artificial intelligence-based project intelligent consulting system and method. Background Art

[0002] Artificial Intelligence (AI) is a branch of computer science that aims to enable machines to perform tasks that usually require human intelligence, including learning, reasoning, problem solving, language understanding, visual recognition, and decision-making. Simply put, the goal of artificial intelligence is to enable machines to think, understand, and act like humans, and even surpass human capabilities in some aspects.

[0003] Traditional systems usually rely on simple rules or keyword-based matching. These methods are usually static and lack a deep understanding of users' personalized needs. Therefore, the recommendation results are often not accurate enough and cannot fully meet users' specific needs. Traditional systems rely on user ratings and comments to recommend items, but these ratings are often affected by false ratings, malicious comments or distorted information. Since there is no effective mechanism to distinguish between false ratings and real ratings, the recommendation system is easily disturbed by these inaccurate information. Traditional systems often rely on fixed rating models or simple rules and lack dynamic adjustment capabilities. Rating updates are usually done manually, requiring human intervention, and the update frequency is low, resulting in the system being unable to respond to changes in the market and user behavior in a timely manner. Summary of the invention

[0004] The technical problem to be solved by the present invention is to overcome the shortcomings of the above-mentioned prior art and provide a project intelligent consulting system and method based on artificial intelligence.

[0005] The technical solution adopted to solve the above technical problems is: an artificial intelligence-based project intelligent consulting system, including an information collection unit, a consulting matching unit, a project scoring unit, a feature extraction unit, a project classification unit, a project recommendation unit and a scoring update unit, specifically: The information collection unit is used to obtain the project consulting information of the target consulting user based on a preset project consulting template, and transmit the project consulting information of the target consulting user to the consulting matching unit; The consulting matching unit is used to receive the project consulting information of the target consulting user transmitted by the information collection unit, match the project consulting information of the target consulting user with a preset project library to obtain a set of candidate projects that match the project consulting information of the target consulting user, and transmit the set of candidate projects to the project scoring unit; The project scoring unit is used to receive the candidate project set transmitted by the consulting matching unit, and obtain historical scoring information of each candidate project in the candidate project set, wherein the historical scoring information includes historical scoring and scoring users, and transmit the historical scoring information of each candidate project in the candidate project set to the feature extraction unit.

[0006] Preferably, the feature extraction unit is used to receive the historical scoring information of each candidate item in the candidate item set transmitted by the item scoring unit, and perform feature extraction on the historical scoring information of each candidate item in the candidate item set to obtain a scoring feature set for each candidate item in the candidate item set, and transmit the scoring feature set for each candidate item in the candidate item set to the item classification unit.

[0007] Preferably, the item classification unit is used to receive the scoring feature set of each candidate item in the candidate item set transmitted by the feature extraction unit, and classify the scoring feature set of each candidate item in the candidate item set based on a pre-trained scoring classification model to obtain a scoring type label corresponding to each candidate item in the candidate item set, wherein the scoring type label includes a real scoring label and a false scoring label, and the candidate item set is divided into a real scoring item set and a false scoring item set based on the item type label, the real scoring item set is transmitted to the item recommendation unit, and the real scoring item set and the false scoring item set are transmitted to the scoring update unit.

[0008] Preferably, the item recommendation unit is used to receive the real-rated item set transmitted by the item classification unit, and recommend items to the target consulting user based on the real-rated item set to obtain an item recommendation list corresponding to the target consulting user.

[0009] Preferably, the scoring update unit is used to receive the real scoring item set and the false scoring item set transmitted by the item classification unit, and match the false scoring item set with the real scoring item set based on a similarity model to obtain real scoring items that match each false scoring item in the false scoring item set, and update the score of each false scoring item in the false scoring item set based on the score of the real scoring item.

[0010] Preferably, feature extraction is performed on the historical scoring information of each candidate item in the candidate item set to obtain a scoring feature set for each candidate item in the candidate item set, including: Calculating the item popularity of each candidate item in the candidate item set based on the historical rating information; Calculating the project novelty of each candidate project in the candidate project set based on the historical scoring information; Sorting the candidate item set according to the item popularity, and dividing the candidate item set into a popularity set and a non-popularity set based on cross-validation; sorting the candidate item set according to the item novelty, and dividing the candidate item set into a novel set and a non-novel set based on cross-validation; Calculate the heat distribution of each heat item in the heat set based on the historical rating information; Calculate the non-thermal distribution of each non-thermal item in the non-thermal set based on the historical rating information; Calculating a novelty distribution of each novel item in the novel set based on the historical scoring information; Calculating a non-novel distribution of each non-novel item in the non-novel set based on the historical rating information; The heat distribution, the non-heat distribution, the novelty distribution and the non-heat distribution are combined to obtain a scoring feature set for each candidate item in the candidate item set.

[0011] Preferably, the calculation formula of the project popularity is as follows: ; in, Represents all rating users' evaluation of candidate items The number of ratings, Indicates the rating user For candidate projects The rating value of ,but , otherwise, , Represents the set of all rating users; The calculation formula of the project novelty is as follows: ; in, Represents all rated user rated candidate items Dissimilarity to other candidates with different ratings, express The total number of items in Indicates the rating user The number of candidate items for scoring, Indicates candidate items With candidate projects The similarity between The calculation formula of the heat distribution is as follows: ; in, Represents candidate items in the heat set The ratio of the heat to the total popularity of the heat set, and , Represents a heat set; The calculation formula of the non-thermal distribution is as follows: ; in, Represents candidate items in a non-hot collection The ratio of the popularity of to the total popularity of the non-hot collection, Represents a non-thermal collection.

[0012] Preferably, the calculation formula of the novel distribution is as follows: ; in, Represents candidate items in the novel set The ratio of the novelty of to the total novelty of the novel set, and , represents a novel set; The calculation formula of the non-novel distribution is as follows: ; in, Represents candidate items in the non-novel set The ratio of the novelty of to the total novelty of the non-novel set, represents a non-novel set.

[0013] Preferably, the scoring classification model adopts an improved support vector machine model, and the penalty parameters and kernel function parameters of the support vector machine model are optimized based on the gray wolf optimization algorithm to obtain the scoring classification model, wherein the penalty parameters and kernel function parameters of the support vector machine model are optimized based on the gray wolf optimization algorithm to obtain the scoring classification model, including: Randomly initializing individuals of the gray wolf population, wherein the positions of the individuals of the gray wolf population represent a parameter combination of a penalty parameter and a kernel function parameter of the support vector machine model; The position of the individual of the gray wolf population is updated to obtain a new position of the individual of the gray wolf population. The new position update formula is as follows: ; in, Indicates The position at the iteration, Indicates the location of the prey. represents the first random parameter, and , represents the nonlinear convergence factor, express The first random number in the interval, represents the distance between the prey and the wolf pack, and express, Indicates The position at the iteration, represents the second random parameter, and , express a second random number in the interval; Calculate the new fitness of the gray wolf population individual at the new position, and compare it with the optimal fitness of the previous iteration. If the new fitness is greater than the optimal fitness, replace the new fitness with the optimal fitness, and retain the new position of the gray wolf population individual; If the number of iterations is greater than the preset iteration threshold, the optimization is terminated to obtain the optimal combination of penalty parameters and kernel function parameters. Otherwise, the above operation is repeated.

[0014] Preferably, recommending items to the target consulting user based on the set of real rated items to obtain a project recommendation list corresponding to the target consulting user includes: Obtaining a set of rating users corresponding to the set of real rating items; Calculating the user similarity between the target consulting user and the set of scoring users; Acquire a K nearest neighbor set of the target consulting user based on the user similarity; Calculate the actual score item deviation based on the actual score corresponding to the K nearest neighbor set of the target consulting user; Calculating the predicted score of the target consulting user for the real scoring item based on the real scoring item deviation; The set of real-rated items is sorted according to the predicted scores, and a preset number of real-rated items in the set of real-rated items are used as recommended items, so as to obtain a project recommendation list corresponding to the target consulting user.

[0015] Preferably, the calculation formula of the user similarity is as follows: ; in, Indicates the rating user and target consulting users In the category The similarity on Indicates the rating user and target consulting users Also rated in the category A collection of projects, Indicates user To belong to The average of the ratings of the items in Indicates the target consulting user To belong to The average of the ratings of the items in ; The calculation formula of the true scoring item deviation is as follows: ; in, Indicates belonging to the same category Project and Projects The average deviation of Indicates that the same category Project and Projects The set of users who rated at the same time, Indicates that they belong to the same category Next and target consultation user The most similar set of top K nearest neighbors; The calculation formula of the prediction score is as follows: ; in, Indicates the target consulting user About Project The prediction score of Indicates that the item In addition, target consulting users Rated in the category A collection of items.

[0016] The technical solution adopted to solve the above technical problems is: an artificial intelligence-based project intelligent consulting method, which is applicable to the artificial intelligence-based project intelligent consulting system, including: Acquiring project consulting information of a target consulting user based on a preset project consulting template, and matching the project consulting information of the target consulting user with a preset project library to obtain a set of candidate projects that match the project consulting information of the target consulting user; Obtaining historical rating information of each candidate item in the candidate item set, wherein the historical rating information includes historical ratings and rating users; Performing feature extraction on the historical scoring information of each candidate item in the candidate item set to obtain a scoring feature set for each candidate item in the candidate item set; Based on a pre-trained scoring classification model, a scoring feature set of each candidate item in the candidate item set is classified to obtain a scoring type label corresponding to each candidate item in the candidate item set, wherein the scoring type label includes a real scoring label and a false scoring label, and the candidate item set is divided into a real scoring item set and a false scoring item set based on the item type label; Recommending items to the target consulting user based on the set of real rated items to obtain a project recommendation list corresponding to the target consulting user; The false-rated item set and the real-rated item set are matched based on a similarity model to obtain real-rated items that match each false-rated item in the false-rated item set, and the score of each false-rated item in the false-rated item set is updated based on the score of the real-rated item.

[0017] The beneficial effects of the present invention are as follows: (1) The present invention obtains the user's specific consulting needs through the information collection unit, and matches them with the preset project library through the consulting matching unit. The system can provide the most relevant candidate project set according to the user's needs. Then, through the historical scoring information and scoring feature extraction of the project scoring unit and the feature extraction unit, the recommended projects are more personalized and accurate, thereby improving the user experience; (2) The present invention classifies the scoring information of the candidate projects through the project classification unit, identifies the falsely rated projects, and removes them from the recommendation list. This classification method can effectively reduce the interference of false scoring or distorted information on the recommendation results, ensuring that the projects recommended to the user are The purpose is to base recommendations on projects with real and valid ratings, thereby improving the reliability and trust of the recommendation system; (3) The present invention uses a rating update unit to match false ratings with real rating projects based on a similarity model and update the ratings of false rating projects, which can effectively and dynamically adjust the project rating system and further improve the accuracy of the ratings. This allows the system to automatically optimize and update the rating model at any time based on newly collected information, maintaining the real-time and accuracy of the system, and adopts a multi-level automated processing flow, reducing the need for manual intervention and improving efficiency and processing speed. Through artificial intelligence algorithms, the system can continue to learn and improve, so that each recommendation is more in line with user needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of the system architecture of the overall system in an embodiment of the present invention; Figure 2 A schematic flow chart of the steps of an overall method in an embodiment of the present invention.

[0019] Figure numerals: 1. Information collection unit; 2. Consultation matching unit; 3. Project scoring unit; 4. Feature extraction unit; 5. Project classification unit; 6. Project recommendation unit; 7. Scoring update unit. DETAILED DESCRIPTION

[0020] Embodiment 1, as Figure 1 As shown, the project intelligent consulting system based on artificial intelligence proposed by the present invention includes an information collection unit 1, a consulting matching unit 2, a project scoring unit 3, a feature extraction unit 4, a project classification unit 5, a project recommendation unit 6 and a scoring update unit 7. Specifically: The information collection unit 1 is used to obtain the project consulting information of the target consulting user based on a preset project consulting template, and transmit the project consulting information of the target consulting user to the consulting matching unit 2; The consulting matching unit 2 is used to receive the project consulting information of the target consulting user transmitted by the information collection unit 1, match the project consulting information of the target consulting user with a preset project library to obtain a set of candidate projects that match the project consulting information of the target consulting user, and transmit the set of candidate projects to the project scoring unit 3; The project scoring unit 3 is used to receive the candidate project set transmitted by the consulting matching unit 2, and obtain the historical scoring information of each candidate project in the candidate project set, wherein the historical scoring information includes the historical scoring and the scoring user, and transmit the historical scoring information of each candidate project in the candidate project set to the feature extraction unit 4.

[0021] In the present invention, target consulting users refer to those users who seek project advice, consultation or matching, usually enterprises or individuals; project consulting information refers to the needs, conditions or expectations submitted by users regarding a certain project, including project type, budget, goals, time frame, resource requirements, etc.; a project library is a collection of multiple project data, usually including project names, descriptions, budgets, time limits, historical execution status and other information, and the project library is usually a database for storing and managing projects in the system; a candidate project set is a set of projects screened by a matching algorithm, and these projects meet the needs of target consulting users or meet matching criteria. In short, these projects are potential candidates for user needs; historical rating information refers to the scores given by rating users when rating a certain project in the past.

[0022] In an optional embodiment, the feature extraction unit 4 is used to receive the historical scoring information of each candidate item in the candidate item set transmitted by the item scoring unit 3, and perform feature extraction on the historical scoring information of each candidate item in the candidate item set to obtain a scoring feature set for each candidate item in the candidate item set, and transmit the scoring feature set for each candidate item in the candidate item set to the item classification unit 5.

[0023] It should be noted that the scoring feature set refers to a set of all scoring-related features extracted from historical scoring information by the feature extraction unit.

[0024] In an optional embodiment, the item classification unit 5 is used to receive the scoring feature set of each candidate item in the candidate item set transmitted by the feature extraction unit 4, and classify the scoring feature set of each candidate item in the candidate item set based on a pre-trained scoring classification model to obtain a scoring type label corresponding to each candidate item in the candidate item set, wherein the scoring type label includes a real scoring label and a false scoring label, and the candidate item set is divided into a real scoring item set and a false scoring item set based on the item type label, and the real scoring item set is transmitted to the item recommendation unit 6, and the real scoring item set and the false scoring item set are transmitted to the scoring update unit 7.

[0025] It should be noted that the rating classification model is a machine learning model that aims to classify ratings based on the rating features of the project. The model is usually trained by analyzing a large amount of historical rating data and learns to distinguish different types of ratings. The goal of the classification model is to predict whether the rating of a project belongs to a specific category, such as "real rating" or "fake rating"; the rating type label refers to the label after classifying the rating of a project. According to the output of the rating classification model, there are two types of rating type labels, one is the real rating label, which indicates that the rating of the project comes from real users and the rating reflects the true quality of the project, and the other is the fake rating label, which indicates that the rating of the project may come from fake or manipulated rating behavior.

[0026] In an optional embodiment, the project recommendation unit 6 is used to receive the real scoring project set transmitted by the project classification unit 5, and recommend projects to the target consulting user based on the real scoring project set to obtain a project recommendation list corresponding to the target consulting user.

[0027] In an optional embodiment, the scoring update unit 7 is used to receive the real scoring item set and the false scoring item set transmitted by the item classification unit 5, and match the false scoring item set with the real scoring item set based on a similarity model to obtain real scoring items that match each false scoring item in the false scoring item set, and update the score of each false scoring item in the false scoring item set based on the score of the real scoring item.

[0028] It should be noted that the similarity model is an algorithm used to measure the similarity between two objects. It determines the degree of similarity by comparing the features of the objects. Similarity models are often used in recommendation systems to discover the similarities between users or projects. Matching refers to pairing or associating false-rated items with real-rated items. In this problem, the purpose of matching is to find real-rated items that are similar to false-rated items. Rating update refers to adjusting the rating of an item according to certain rules or algorithms. The rating of each item in the set of false-rated items will be updated according to the rating of similar real-rated items. Assuming that a false-rated item is artificially manipulated during rating, its rating may not be accurate. By comparing it with the matched real-rated item, the rating of the false-rated item can be updated to make it more consistent with the rating trend of the real item, thereby improving the credibility of the rating.

[0029] Embodiment 2, an artificial intelligence-based project intelligent consulting system proposed by the present invention, compared with embodiment 1, this embodiment further includes: extracting features from the historical scoring information of each candidate project in the candidate project set to obtain a scoring feature set for each candidate project in the candidate project set, including: Calculate the project popularity of each candidate project in the candidate project set based on historical rating information; Calculate the project novelty of each candidate project in the candidate project set based on historical rating information; Sort the candidate project set according to the project popularity, and divide the candidate project set into a hot set and a non-hot set based on cross-validation; Sort the candidate project set according to the project novelty, and divide the candidate project set into a novel set and a non-novel set based on cross-validation; Calculate the heat distribution of each hot item in the heat set based on the historical rating information; Calculate the non-hotness distribution of each non-hotness item in the non-hotness set based on the historical rating information; Calculate the novelty distribution of each novel item in the novel set based on the historical rating information; Calculate the non-novel distribution of each non-novel item in the non-novel set based on the historical rating information; The heat distribution, the non-heat distribution, the novelty distribution, and the non-heat distribution are combined to obtain a scoring feature set for each candidate item in the candidate item set.

[0030] In this embodiment, the popularity of an item is an indicator that measures the popularity of an item in historical rating data. Generally speaking, high-popularity items refer to those that are rated or visited by a large number of users, and may represent the mainstream preferences of current users; the novelty of an item measures the "freshness" of an item in the recommendation list to the user, that is, whether the item is a unique or less-rated item that the user has never come into contact with. High-novelty items are usually newly released, unpopular, or items that the user has never come into contact with; the hot set refers to a set of items that are considered to be popular and have a large number of ratings in a certain candidate item set after being sorted according to historical rating information. Usually, these items have high historical ratings and may have received widespread attention and interaction; the non-hot set refers to those items with low popularity and a large number of ratings. A smaller set of projects, usually these projects are not popular and may be unpopular projects; novel set refers to the set of projects with higher novelty in the candidate project set after being sorted based on the novelty of the projects; non-novel set refers to the set of projects with lower novelty in the candidate project set after being sorted based on the novelty of the projects; heat distribution refers to the performance of the score distribution of a certain project in the heat set in the historical scoring information; non-heat distribution refers to the performance of the score distribution of a certain project in the non-heat set in the historical scoring information; novel distribution refers to the performance of the score distribution of a certain project in the novel set in the historical scoring information; non-novel distribution refers to the performance of the score distribution of a certain project in the non-novel set in the historical scoring information.

[0031] In an optional embodiment, the calculation formula of the project popularity is as follows: ; in, Represents all rating users' responses to candidate items The number of ratings, Indicates the rating user For candidate projects The rating value of ,but , otherwise, , Represents the set of all rating users; The calculation formula of project novelty is as follows: ; in, Represents all rated user rated candidate items Dissimilarity to other candidates with different ratings, express The total number of items in Indicates the rating user The number of candidate items for scoring, Indicates candidate items With candidate projects The similarity between The calculation formula of heat distribution is as follows: ; in, Represents candidate items in the heat set The ratio of the heat to the total popularity of the heat set, and , Represents a heat set; The calculation formula of non-thermal distribution is as follows: ; in, Represents candidate items in a non-hot collection The ratio of the popularity of to the total popularity of the non-hot collection, Represents a non-thermal collection.

[0032] In an optional embodiment, the calculation formula of the novel distribution is as follows: ; in, Represents candidate items in the novel set The ratio of the novelty of to the total novelty of the novel set, and , represents a novel set; The non-novel distribution is calculated as follows: ; in, Represents candidate items in the non-novel set The ratio of the novelty of to the total novelty of the non-novel set, represents a non-novel set.

[0033] In an optional embodiment, the scoring classification model adopts an improved support vector machine model, and the penalty parameters and kernel function parameters of the support vector machine model are optimized based on the gray wolf optimization algorithm to obtain the scoring classification model, wherein the penalty parameters and kernel function parameters of the support vector machine model are optimized based on the gray wolf optimization algorithm to obtain the scoring classification model, including: The individuals of the gray wolf population are randomly initialized, and the positions of the individuals of the gray wolf population represent the parameter combination of the penalty parameter and the kernel function parameter of the support vector machine model; The positions of the individuals in the gray wolf population are updated to obtain the new positions of the individuals in the gray wolf population. The formula for updating the new positions is as follows: ; in, Indicates The position at the iteration, Indicates the location of the prey. represents the first random parameter, and , represents the nonlinear convergence factor, express The first random number in the interval, represents the distance between the prey and the wolf pack, and express, Indicates The position at the iteration, represents the second random parameter, and , express a second random number in the interval; Calculate the new fitness of the gray wolf population individual at the new position and compare it with the optimal fitness of the previous iteration. If the new fitness is greater than the optimal fitness, replace the new fitness with the optimal fitness and keep the new position of the gray wolf population individual. If the number of iterations is greater than the preset iteration threshold, the optimization is terminated to obtain the optimal combination of penalty parameters and kernel function parameters. Otherwise, the above operation is repeated.

[0034] It should be noted that support vector machine is a supervised learning algorithm for classification and regression tasks. It aims to classify data by maximizing the interval between categories (i.e., decision boundary). SVM uses one or more kernel functions to map data to high-dimensional space, making it linearly separable. The penalty parameter, also called the constant parameter, is a hyperparameter in SVM. It controls the model's tolerance for misclassification. A larger C value will make the model fit the training data more strictly, which may lead to overfitting, while a smaller C value may cause the model to underfit. The kernel function is used to map data to a higher dimension in order to achieve linear segmentation in high-dimensional space. Common kernel functions include radial basis function (RBF) kernel, polynomial kernel, etc. The parameters of the kernel function (such as the γ parameter in the RBF kernel) affect the structure of the high-dimensional space after mapping. The gray wolf optimization algorithm (GWO) is a heuristic optimization algorithm that simulates the hunting behavior of gray wolf groups. GWO finds the optimal solution to the problem by simulating the hunting process between gray wolves (divided into leading wolves, following wolves, and prey). This algorithm is often used to solve complex optimization problems. In optimization algorithms, fitness is a measure of the quality of an individual (or solution). Usually, the fitness function is defined according to the goal of the problem and represents the quality of a solution. In gray wolf optimization, fitness reflects the quality of the individual gray wolf solution and is usually related to the objective function (such as classification accuracy).

[0035] In an optional embodiment, project recommendations are made to a target consulting user based on the real rated project set to obtain a project recommendation list corresponding to the target consulting user, including: Get the set of rated users corresponding to the set of real rated items; Calculate the user similarity between the target consulting user and the set of rated users; Obtain the K nearest neighbor set of the target consulting user based on user similarity; Calculate the actual rating item deviation based on the actual rating corresponding to the K nearest neighbor set of the target consulting user; Calculate the target consulting user's predicted rating for the real rating item based on the deviation of the real rating item; The real-rated item set is sorted according to the predicted scores, and a preset number of real-rated items in the real-rated item set are used as recommended items to obtain a project recommendation list corresponding to the target consulting user.

[0036] It should be noted that the K-nearest neighbor algorithm (KNN) is a commonly used similarity-based algorithm. The K-nearest neighbor set refers to the set of K users with the highest similarity to the target consulting user. The users in this set are considered to have reference value for recommendations to the target user; deviation refers to the deviation between a user's score for an item and the score of the item given to the item by other users in the user's K-nearest neighbor set; predicted score refers to estimating the possible score of the user for an item based on the behavioral data of the target consulting user and other users in his K-nearest neighbor set; sorting refers to sorting the predicted scores, usually by the score, with the items most likely to be liked by the target user placed in front; the recommended item list refers to a set of personalized recommended items generated based on the user's interests and predicted scores.

[0037] In an optional embodiment, the calculation formula of user similarity is as follows: ; in, Indicates the rating user and target consulting users In the category The similarity on Indicates the rating user and target consulting users Also rated in the category A collection of projects, Indicates user To belong to The average of the ratings of the items in Indicates the target consulting user To belong to The average of the ratings of the items in ; The calculation formula for the true rating item deviation is as follows: ; in, Indicates belonging to the same category Project and Projects The average deviation of Indicates that the same category Project and Projects The set of users who rated at the same time, Indicates that they belong to the same category Next and target consultation user The most similar set of top K nearest neighbors; The calculation formula for the prediction score is as follows: ; in, Indicates the target consulting user About Project The prediction score of Indicates that the item In addition, target consulting users Rated in the category A collection of items.

[0038] Embodiment three, as Figure 2 As shown, the present invention proposes an artificial intelligence-based project intelligent consulting method, which is applicable to the artificial intelligence-based project intelligent consulting system, including: S1. Acquire project consulting information of a target consulting user based on a preset project consulting template, and match the project consulting information of the target consulting user with a preset project library to obtain a set of candidate projects that match the project consulting information of the target consulting user; S2. Obtain historical rating information of each candidate item in the candidate item set, wherein the historical rating information includes historical ratings and rating users; S3, extracting features from the historical scoring information of each candidate item in the candidate item set to obtain a scoring feature set for each candidate item in the candidate item set; S4. Classifying the scoring feature set of each candidate item in the candidate item set based on the pre-trained scoring classification model to obtain a scoring type label corresponding to each candidate item in the candidate item set, wherein the scoring type label includes a real scoring label and a false scoring label, and dividing the candidate item set into a real scoring item set and a false scoring item set based on the item type label; S5. Recommending projects to the target consulting user based on the real scoring project set to obtain a project recommendation list corresponding to the target consulting user; S6. Match the false-rated item set and the real-rated item set based on the similarity model to obtain the real-rated item that matches each false-rated item in the false-rated item set, and update the score of each false-rated item in the false-rated item set based on the score of the real-rated item.

[0039] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited thereto, and various changes can be made within the knowledge scope of technicians in the relevant technical field without departing from the purpose of the present invention.

Claims

1. An artificial intelligence-based project intelligent consulting system, characterized in that: It includes an information collection unit (1), a consultation matching unit (2), a project scoring unit (3), a feature extraction unit (4), a project classification unit (5), a project recommendation unit (6) and a scoring update unit (7), specifically: The information collection unit (1) is used to obtain the project consulting information of the target consulting user based on a preset project consulting template, and transmit the project consulting information of the target consulting user to the consulting matching unit (2); The consultation matching unit (2) is used to receive the project consultation information of the target consultation user transmitted by the information collection unit (1), match the project consultation information of the target consultation user with a preset project library to obtain a candidate project set that matches the project consultation information of the target consultation user, and transmit the candidate project set to the project scoring unit (3); The project scoring unit (3) is used to receive the candidate project set transmitted by the consulting matching unit (2), and obtain the historical scoring information of each candidate project in the candidate project set, wherein the historical scoring information includes the historical scoring and the scoring user, and transmit the historical scoring information of each candidate project in the candidate project set to the feature extraction unit (4).

2. According to claim 1, the project intelligent consulting system based on artificial intelligence is characterized in that: The feature extraction unit (4) is used to receive the historical scoring information of each candidate item in the candidate item set transmitted by the item scoring unit (3), and perform feature extraction on the historical scoring information of each candidate item in the candidate item set to obtain a scoring feature set for each candidate item in the candidate item set, and transmit the scoring feature set for each candidate item in the candidate item set to the item classification unit (5); The item classification unit (5) is used to receive the scoring feature set of each candidate item in the candidate item set transmitted by the feature extraction unit (4), and classify the scoring feature set of each candidate item in the candidate item set based on a pre-trained scoring classification model to obtain a scoring type label corresponding to each candidate item in the candidate item set, wherein the scoring type label includes a real scoring label and a false scoring label, and the candidate item set is divided into a real scoring item set and a false scoring item set based on the item type label, and the real scoring item set is transmitted to the item recommendation unit (6), and the real scoring item set and the false scoring item set are transmitted to the scoring update unit (7).

3. According to claim 2, the project intelligent consulting system based on artificial intelligence is characterized in that: The project recommendation unit (6) is used to receive the real scoring project set transmitted by the project classification unit (5), and recommend projects to the target consulting user based on the real scoring project set, so as to obtain a project recommendation list corresponding to the target consulting user; The scoring update unit (7) is used to receive the real scoring item set and the false scoring item set transmitted by the item classification unit (5), and match the false scoring item set with the real scoring item set based on a similarity model to obtain a real scoring item that matches each false scoring item in the false scoring item set, and update the score of each false scoring item in the false scoring item set based on the score of the real scoring item.

4. According to claim 2, the project intelligent consulting system based on artificial intelligence is characterized in that: Performing feature extraction on the historical scoring information of each candidate item in the candidate item set to obtain a scoring feature set for each candidate item in the candidate item set, including: Calculating the item popularity of each candidate item in the candidate item set based on the historical rating information; Calculating the project novelty of each candidate project in the candidate project set based on the historical scoring information; Sorting the candidate item set according to the item popularity, and dividing the candidate item set into a popularity set and a non-popularity set based on cross-validation; sorting the candidate item set according to the item novelty, and dividing the candidate item set into a novel set and a non-novel set based on cross-validation; Calculate the heat distribution of each heat item in the heat set based on the historical rating information; Calculate the non-thermal distribution of each non-thermal item in the non-thermal set based on the historical rating information; Calculating a novelty distribution of each novel item in the novel set based on the historical scoring information; Calculating a non-novel distribution of each non-novel item in the non-novel set based on the historical rating information; The heat distribution, the non-heat distribution, the novelty distribution and the non-heat distribution are combined to obtain a scoring feature set for each candidate item in the candidate item set.

5. The project intelligent consulting system based on artificial intelligence according to claim 4 is characterized in that: The calculation formula of the project heat is as follows: ； in, Represents all rating users' responses to candidate items The number of ratings, Indicates the rating user For candidate projects The rating value of ,but , otherwise, , Represents the set of all rating users; The calculation formula of the project novelty is as follows: ； in, Represents all rated user rated candidate items Dissimilarity to other candidates with different ratings, express The total number of items in Indicates the rating user The number of candidate items for scoring, Indicates candidate items With candidate projects The similarity between The calculation formula of the heat distribution is as follows: ； in, Represents candidate items in the heat set The ratio of the heat to the total popularity of the heat set, and , Represents a heat set; The calculation formula of the non-thermal distribution is as follows: ； in, Represents candidate items in a non-hot collection The ratio of the popularity of to the total popularity of the non-hot collection, Represents a non-thermal collection.

6. The project intelligent consulting system based on artificial intelligence according to claim 5 is characterized in that: The calculation formula of the novel distribution is as follows: ； in, Represents candidate items in the novel set The ratio of the novelty of to the total novelty of the novel set, and , represents a novel set; The calculation formula of the non-novel distribution is as follows: ； in, Represents candidate items in the non-novel set The ratio of the novelty of to the total novelty of the non-novel set, represents a non-novel set.

7. The project intelligent consulting system based on artificial intelligence according to claim 2 is characterized in that: The scoring classification model adopts an improved support vector machine model, and optimizes the penalty parameters and kernel function parameters of the support vector machine model based on the gray wolf optimization algorithm to obtain the scoring classification model, wherein the penalty parameters and kernel function parameters of the support vector machine model are optimized based on the gray wolf optimization algorithm to obtain the scoring classification model, including: Randomly initializing individuals of the gray wolf population, wherein the positions of the individuals of the gray wolf population represent a parameter combination of a penalty parameter and a kernel function parameter of the support vector machine model; The position of the individual of the gray wolf population is updated to obtain a new position of the individual of the gray wolf population. The new position update formula is as follows: ； in, Indicates The position at the iteration, Indicates the location of the prey. represents the first random parameter, and , represents the nonlinear convergence factor, express The first random number in the interval, represents the distance between the prey and the wolf pack, and express, Indicates The position at the iteration, represents the second random parameter, and , express a second random number in the interval; Calculate the new fitness of the gray wolf population individual at the new position, and compare it with the optimal fitness of the previous iteration. If the new fitness is greater than the optimal fitness, replace the new fitness with the optimal fitness, and retain the new position of the gray wolf population individual; If the number of iterations is greater than the preset iteration threshold, the optimization is terminated to obtain the optimal combination of penalty parameters and kernel function parameters. Otherwise, the above operation is repeated.

8. The project intelligent consulting system based on artificial intelligence according to claim 3 is characterized in that: Recommending items to the target consulting user based on the set of real rated items to obtain a project recommendation list corresponding to the target consulting user includes: Obtaining a set of rating users corresponding to the set of real rating items; Calculating the user similarity between the target consulting user and the set of scoring users; Acquire a K nearest neighbor set of the target consulting user based on the user similarity; Calculate the actual score item deviation based on the actual score corresponding to the K nearest neighbor set of the target consulting user; Calculating the predicted score of the target consulting user for the real scoring item based on the real scoring item deviation; The set of real-rated items is sorted according to the predicted scores, and a preset number of real-rated items in the set of real-rated items are used as recommended items, so as to obtain a project recommendation list corresponding to the target consulting user.

9. The project intelligent consulting system based on artificial intelligence according to claim 8 is characterized in that: The calculation formula of the user similarity is as follows: ； in, Indicates the rating user and target consulting users In the category The similarity on Indicates the rating user and target consulting users Also rated in the category A collection of projects, Indicates user To belong to The average of the ratings of the items in Indicates the target consulting user To belong to The average of the ratings of the items in ; The calculation formula of the true scoring item deviation is as follows: ； in, Indicates belonging to the same category Project and Projects The average deviation of Indicates that the same category Project and Projects The set of users who rated at the same time, Indicates that they belong to the same category Next and target consultation user The most similar set of top K nearest neighbors; The calculation formula of the prediction score is as follows: ； in, Indicates the target consulting user About Project The prediction score of Indicates that the item In addition, target consulting users Rated in the category A collection of items.

10. An artificial intelligence-based project intelligent consulting method, which is applicable to an artificial intelligence-based project intelligent consulting system according to any one of claims 1 to 9, characterized in that: include: Acquiring project consulting information of a target consulting user based on a preset project consulting template, and matching the project consulting information of the target consulting user with a preset project library to obtain a set of candidate projects that match the project consulting information of the target consulting user; Obtaining historical rating information of each candidate item in the candidate item set, wherein the historical rating information includes historical ratings and rating users; Performing feature extraction on the historical scoring information of each candidate item in the candidate item set to obtain a scoring feature set for each candidate item in the candidate item set; Based on a pre-trained scoring classification model, a scoring feature set of each candidate item in the candidate item set is classified to obtain a scoring type label corresponding to each candidate item in the candidate item set, wherein the scoring type label includes a real scoring label and a false scoring label, and the candidate item set is divided into a real scoring item set and a false scoring item set based on the item type label; Recommending items to the target consulting user based on the set of real rated items to obtain a project recommendation list corresponding to the target consulting user; The false-rated item set and the real-rated item set are matched based on a similarity model to obtain real-rated items that match each false-rated item in the false-rated item set, and the score of each false-rated item in the false-rated item set is updated based on the score of the real-rated item.

Images