Personalized shopping guide method and system based on artificial intelligence

Abstract

The invention relates to the technical field of preference prediction, and discloses a personalized shopping guide method and system based on artificial intelligence. The method comprises the following steps: acquiring basic information of a user and historical consumption commodity attributes; according to the basic information of the user and historical consumption commodity attributes, preference prediction is carried out to obtain a user preference score; constructing a decision matrix according to the user preference score, and calculating to obtain a weighted score matrix; the maximum value in the weighted score matrix is selected to be determined as a positive ideal score, and the minimum value in the weighted score matrix is selected to be determined as a negative ideal score; respectively calculating the distance between the positive ideal score and the standardized score and the distance between the negative ideal score and the standardized score to obtain a positive ideal distance and a negative ideal distance; and calculating preference score closeness according to the positive ideal distance and the negative ideal distance, and selecting an alternative scheme corresponding to the maximum preference score closeness as a recommendation result. According to the method, more accurate commodity recommendation can be provided for specific preferences and purchase intentions of the user.

Description

Technical Field

[0001] The present invention relates to the technical field of preference prediction, and in particular to a personalized shopping guide method and system based on artificial intelligence. Background Art

[0002] With the rise of artificial intelligence, personalized smart shopping guides have gradually replaced traditional shopping guides, saving consumers’ time and improving shopping experience. Personalized smart shopping guides can not only shorten the distance between consumers and products, but also provide consumers with the most ideal products based on understanding consumers’ purchase intentions and product characteristics, thus realizing a user-centered shopping method.

[0003] Currently, shopping guide services are mainly provided on mobile shopping apps based on product sales history, merchant comprehensive evaluation, consumer purchase history, etc., and are displayed directly through sorting, etc. Although this method can also assist consumers in choosing the right products and solve the problem of consumers' shopping time, it still cannot fully meet consumers' personalized and intelligent needs.

[0004] In summary, the current screening method of e-commerce apps is mainly based on product sales, merchant reviews and historical purchase records. In the absence of personalized preferences, this method results in recommended products that cannot fully match user needs. Summary of the invention

[0005] The present invention provides a personalized shopping guide method and system based on artificial intelligence, which can focus on the personalized needs of users and provide more accurate commodity recommendations based on the specific preferences and purchasing intentions of users.

[0006] In a first aspect, in order to solve the above technical problems, the present invention provides a personalized shopping guide method based on artificial intelligence, comprising: obtaining basic information of a user and attributes of historically consumed goods; wherein the basic information of the user includes the gender, age and height of the user, and the attributes of historically consumed goods include the style and brand of the goods;

[0007] According to the basic information of the user and the attributes of the historically consumed goods, preference prediction is performed to obtain a user preference score;

[0008] A decision matrix is ​​constructed according to the user preference scores, and a weighted score matrix is ​​obtained by performing weighted calculation according to the user preference scores and the decision matrix; wherein the decision matrix is ​​composed of the user preference scores, the rows of the decision matrix represent alternatives, and the columns of the decision matrix represent the attributes of the historically consumed goods;

[0009] Select the maximum value in the weighted scoring matrix to determine it as a positive ideal score, and select the minimum value in the weighted scoring matrix to determine it as a negative ideal score;

[0010] Calculating the distances between the positive ideal score, the negative ideal score and the standardized score respectively to obtain a positive ideal distance and a negative ideal distance; wherein the standardized score is calculated from the user preference score;

[0011] The preference score proximity is calculated according to the positive ideal distance and the negative ideal distance, and the preference score proximity is sorted, and the alternative corresponding to the largest preference score proximity is selected as the recommendation result.

[0012] Preferably, performing preference prediction based on the basic information of the user and the attributes of the historically consumed goods to obtain a user preference score includes:

[0013] According to the basic information of the user, commodity prediction is performed based on the k-means clustering algorithm and the Markov chain algorithm to obtain the commodity type;

[0014] According to the attributes of the historically consumed goods and the types of goods, preference prediction is performed based on the k-means clustering algorithm to obtain a user preference score.

[0015] Preferably, performing weighted calculation according to the user preference score and the decision matrix to obtain a weighted score matrix includes:

[0016] According to the user preference scores, a standardized score is calculated to obtain a standardized score matrix; wherein the standardized score matrix is ​​composed of the standardized scores;

[0017] According to the decision matrix, calculating the column average of each column in the decision matrix;

[0018] Calculate the column standard deviation based on the column average and the user preference score in each column;

[0019] According to the column standard deviation, weight calculation is performed to obtain column weight;

[0020] The column weight is multiplied by each element in the normalized score matrix to obtain a weighted score matrix.

[0021] Preferably, the step of calculating the standardized scores according to the user preference scores to obtain a standardized score matrix includes:

[0022] Calculating a standardized score based on the user preference score;

[0023] A standardized scoring matrix is ​​constructed according to the standardized scores; wherein the standardized scoring matrix is ​​composed of the standardized scores.

[0024] Preferably, calculating the standardized score according to the user preference score includes:

[0025] The normalized score is calculated using the following formula:

[0026]

[0027] In the formula, y ij is the standardized score, x ij Score user preference, min(x j ) represents the minimum user preference score of the jth column, mαx(x j ) represents the maximum user preference score of the jth column.

[0028] Preferably, calculating the column standard deviation according to the column average value and the user preference score in each column includes:

[0029] The column standard deviation is calculated by the following formula:

[0030]

[0031] In the formula, σ j is the column standard deviation, x ij Rate user preferences, is the column average and n is the total number of rows.

[0032] Preferably, performing weight calculation according to the column standard deviation to obtain the column weight includes:

[0033] The column weight is calculated by the following formula:

[0034]

[0035] In the formula, w j is the column weight, σ j is the column standard deviation, Represents the sum of the column standard deviations.

[0036] Preferably, the calculating the preference score proximity according to the positive ideal distance and the negative ideal distance comprises:

[0037] The preference score proximity is calculated by the following formula:

[0038]

[0039] In the formula, is the negative ideal distance, is the positive ideal distance, C i Score proximity for preferences.

[0040] In a second aspect, the present invention provides a personalized shopping guide system based on artificial intelligence, comprising:

[0041] A data acquisition module is used to acquire basic user information and attributes of historically consumed goods; wherein the basic user information includes the user's gender, age and height, and the attributes of historically consumed goods include the style and brand of the goods;

[0042] A preference prediction module, used to make preference predictions based on the basic information of the user and the attributes of the historically consumed goods, and obtain a user preference score;

[0043] A matrix construction module, used to construct a decision matrix according to the user preference scores, and perform weighted calculation according to the user preference scores and the decision matrix to obtain a weighted score matrix; wherein the decision matrix is ​​composed of the user preference scores, the rows of the decision matrix represent alternatives, and the columns of the decision matrix represent the attributes of the historical consumption commodities;

[0044] A score determination module, configured to select the maximum value in the weighted score matrix to be determined as a positive ideal score, and select the minimum value in the weighted score matrix to be determined as a negative ideal score;

[0045] A distance calculation module, used to calculate the distance between the positive ideal score, the negative ideal score and the standardized score, respectively, to obtain a positive ideal distance and a negative ideal distance; wherein the standardized score is calculated from the user preference score;

[0046] The recommendation calculation module is used to calculate the preference score proximity according to the positive ideal distance and the negative ideal distance, sort the preference score proximity, and select the alternative corresponding to the largest preference score proximity as the recommendation result.

[0047] In a third aspect, the present invention further provides an electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the personalized shopping guide method based on artificial intelligence as described above is implemented.

[0048] In a fourth aspect, the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned artificial intelligence-based personalized shopping guide methods.

[0049] Compared with the prior art, the present invention has the following beneficial effects: The embodiment of the present invention provides a personalized shopping guide method and system based on artificial intelligence. The method comprises: obtaining basic information of a user and attributes of historically consumed goods; wherein the basic information of the user comprises the gender, age and height of the user, and the attributes of historically consumed goods comprise the style and brand of the goods; performing preference prediction according to the basic information of the user and the attributes of historically consumed goods to obtain a user preference score; constructing a decision matrix according to the user preference score, and performing weighted calculation according to the user preference score and the decision matrix to obtain a weighted score matrix; wherein the decision matrix is ​​composed of the user preference scores, the rows of the decision matrix represent alternatives, and the columns of the decision matrix represent the attributes of historically consumed goods; selecting the maximum value in the weighted score matrix to determine as a positive ideal score, and selecting the minimum value in the weighted score matrix to determine as a negative ideal score; respectively calculating the distances between the positive ideal score, the negative ideal score and the standardized score to obtain a positive ideal distance and a negative ideal distance; wherein the standardized score is calculated from the user preference score; calculating the preference score proximity according to the positive ideal distance and the negative ideal distance, and sorting the preference score proximity, and selecting the alternative corresponding to the largest preference score proximity as a recommendation result.

[0050] In the present invention, the method can obtain the basic information of the user (such as gender, age, height) and the attributes of historical consumption goods (such as the style and brand of the goods), use artificial intelligence algorithms (k-means clustering algorithm and Markov chain algorithm) to predict preferences, and combine the TOPSIS method to perform recommendation calculations. Specifically, the basic information of the user and the attributes of historical consumption goods are obtained, k-means clustering is performed based on the basic information of the user (such as gender, age, height), the correlation between the user and the goods is analyzed, and the type of goods suitable for the user is preliminarily determined. Combined with the historical consumption goods attributes of the user, the purchase behavior transfer probability is analyzed by the Markov chain algorithm, and the user's interest preference is further accurately predicted. A decision matrix containing user preference scores is constructed, and the column standard deviation is calculated by the standardized score matrix to obtain the weight distribution. The TOPSIS method is used to combine the ideal score and the negative ideal score, calculate the proximity of the preference score and sort it, and finally filter out the product recommendation results that best meet the user's needs. In summary, the method can focus on the personalized needs of users and provide more accurate product recommendations for users' specific preferences and purchase intentions. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a flowchart of a personalized shopping guide method based on artificial intelligence provided by the first embodiment of the present invention;

[0052] Figure 2It is a structural diagram of an artificial intelligence-based personalized shopping guide system provided by the second embodiment of the present invention. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] Reference Figure 1 The first embodiment of the present invention provides a personalized shopping guide method based on artificial intelligence, comprising the following steps:

[0055] S11, obtaining basic information of the user and attributes of historically consumed goods.

[0056] S12, performing preference prediction based on the basic information of the user and the attributes of the historically consumed commodities to obtain a user preference score.

[0057] S13, constructing a decision matrix according to the user preference scores, and performing weighted calculation according to the user preference scores and the decision matrix to obtain a weighted score matrix.

[0058] S14, selecting the maximum value in the weighted scoring matrix to determine as the positive ideal score, and selecting the minimum value in the weighted scoring matrix to determine as the negative ideal score.

[0059] S15, respectively calculating the distances between the positive ideal score, the negative ideal score and the standardized score to obtain a positive ideal distance and a negative ideal distance.

[0060] S16, calculating preference score proximity according to the positive ideal distance and the negative ideal distance, sorting the preference score proximity, and selecting the alternative corresponding to the largest preference score proximity as the recommendation result.

[0061] It is worth noting that both personalized shopping guides and personalized recommendations are strategies that optimize shopping or service experiences through technical means with user needs as the core, but their implementation methods and application scenarios are different. Personalized shopping guides are more like simulating the role of offline shopping guides online, providing targeted suggestions based on the needs clearly expressed by users through real-time interaction with users. It usually helps users gradually narrow down their choices through question-and-answer guidance and input information screening. It is suitable for scenarios where the products are complex and users need clear guidance, such as the selection of home appliances, financial products or cosmetics. In contrast, personalized recommendations are more data-driven and do not require users to actively express their needs. Instead, they automatically recommend products or services that may be of interest to users by analyzing their historical behaviors, interest preferences and purchase records. This approach emphasizes the automation and intelligence of the system and is suitable for scenarios where user needs are unclear or potential needs need to be stimulated, such as similar product recommendations on e-commerce platforms or movie recommendations on streaming platforms. The common goal of both is to improve user experience, but personalized shopping guides provide users with clearer guidance through interaction, while personalized recommendations predict user interests through algorithms and provide convenient suggestions.

[0062] In order to facilitate the understanding of the present invention, some preferred embodiments of the present invention are further described below.

[0063] In step S11, basic user information and attributes of historically consumed goods are obtained.

[0064] Preferably, the basic user information includes the user's gender, age and height, and the attributes of historically consumed goods include the style and brand of the goods.

[0065] Specifically, methods for obtaining basic user information may include: retrieving registration information, personal information filled out by users when registering or using the platform, such as gender, age, height, etc.; user behavior analysis, inferring some information by analyzing user operations on the platform (such as selecting product types, browsing time, interest classification, etc.); retrieving social account authorization, when users log in through a third-party account (such as social media), the platform can obtain basic information within the scope of authorization; using form interaction, actively obtaining users' basic characteristics through question-and-answer format during the shopping guide or recommendation process.

[0066] Specifically, the methods for obtaining the attributes of historical consumption goods may include: user transaction records, user's historical order information on the platform, including the brand and style of purchased goods; user browsing records, information on goods that users have browsed or added to the shopping cart but not purchased; user evaluations, user evaluations and preference feedback on purchased goods; platform external data, user purchase data obtained from other platforms through data sharing agreements (data privacy and security must be ensured).

[0067] It is worth noting that, combined with this information, the system can extract user-related data from the database through a specific data collection module or API interface for subsequent preference prediction and recommendation calculations.

[0068] In step S12, preference prediction is performed based on the user basic information and the attributes of the historically consumed goods to obtain a user preference score.

[0069] Preferably, performing preference prediction based on the basic information of the user and the attributes of the historically consumed goods to obtain a user preference score includes:

[0070] According to the basic information of the user, commodity prediction is performed based on the k-means clustering algorithm and the Markov chain algorithm to obtain the commodity type;

[0071] According to the attributes of the historically consumed goods and the types of goods, preference prediction is performed based on the k-means clustering algorithm to obtain a user preference score.

[0072] It is worth mentioning that the k-means clustering algorithm is a machine learning method for dividing data points into different groups (called clusters) according to their similarities. It assigns data points to k clusters through a repeated optimization process, so that the data points in each cluster have the greatest degree of similarity in specific features, while the similarity between different clusters is minimal. The goal of k-means is to find k center points (centroids) so that the distance between each data point and the center point of its cluster is as small as possible. In other words, data points are divided into clusters where their "closest" center points are located. The algorithm flow can be summarized as follows: initialization, randomly select k points as initial cluster centers; assign samples, assign each data point to the nearest cluster center to form k clusters; update the center, calculate the average value of all data points in each cluster, and use the average value as the new cluster center; iteration, repeat steps 2 and 3 until the cluster center is stable or the set number of iterations is reached.

[0073] It is worth noting that the Markov chain algorithm is a mathematical model used to describe the random transition process between states. It is based on a key assumption that "the future state depends only on the current state and has nothing to do with the past state", a property called "no memory". The goal of the Markov chain is to predict possible future states by analyzing the probability of a system transitioning from one state to another. Its components are: a state set, which defines all possible states. For example, in commodity prediction, the state set can be different commodity categories (such as "sports shoes", "sportswear", and "accessories"); state transition probability, the probability of transitioning from one state to another. For example, the probability that a user transfers from buying "sports shoes" to buying "sportswear" is 0.4; a transition matrix, which records the transition probabilities between all states in a tabular form, where rows represent the current state, columns represent the target state, and the sum of the probabilities of each row in the matrix is ​​1. Its working process can be summarized as follows: the system first starts from an initial state (such as the category of commodities that the user recently purchased); according to the transition probability matrix, the system randomly selects the next state. For example, if the current state is "sports shoes", then according to the matrix, the probability that the user may buy "sportswear" next time is 0.5. Through multiple state transitions, the user's possible future purchase path can be predicted. For example, from "sports shoes" to "sportswear" and then to "accessories".

[0074] Specifically, in the present invention, k-means clustering is used to group users. In one achievable manner, users are divided into different groups (such as "young sports users" and "middle-aged business users") by analyzing the gender, age and height of the users. For each user group, the Markov chain constructs a transition probability matrix using historical purchase behavior.

[0075] Specifically, based on k-means clustering, the basic information of users (such as gender, age, height, etc.) is converted into feature vectors. For example, gender is assumed to be a binary variable (0 for male and 1 for female), and age and height are continuous variables. According to the feature vectors, a feature space is constructed. The feature space is divided into k clusters using the k-means algorithm, and each cluster represents a user group. The result of clustering is that each user is assigned to a specific cluster, and the center point of the cluster reflects the common characteristics of the group. For each user in each cluster, their degree of association with a specific type of product is counted. The basic information of the new user is input into the k-means model, the distance from the center point of each cluster is calculated, and the user is assigned to the nearest cluster. The potential product type of the new user is inferred based on the product type preference of the cluster.

[0076] Specifically, based on the Markov chain, the commodity type is regarded as a state set, and according to the common characteristics and statistical results of the user group, a transition probability matrix between commodity types is constructed, in which the rows of the matrix represent the current commodity type and the columns represent the commodity type that may be transferred to. According to the k-means result, the user is assigned to an initial commodity type. According to the transition probability matrix, starting from the initial state, the possible commodity types in the future are simulated. Exemplarily, if the long-term trend is considered, the commodity type distribution of the user may tend to a steady state through multiple transfers. In summary, in the present invention, the prediction process combining the two algorithms includes: based on the k-means preliminary classification, the basic information of the user is input into the k-means model, according to the clustering results, the user is assigned to a specific group, and according to the commodity type preference of the group, a preliminary prediction result is generated; based on the Markov chain dynamic prediction, the commodity type predicted by k-means is used as the initial state, and according to the transition probability matrix of the commodity type, the possible commodity type in the future is predicted, and the results are comprehensively analyzed to generate the final commodity type prediction.

[0077] Specifically, in order to predict preferences and obtain user preference scores based on historical consumption commodity attributes and commodity types through the k-means clustering algorithm, it is necessary to combine the user's consumption records with the commodity types predicted by the system and use clustering methods to analyze the user's overall preference characteristics.

[0078] Specifically, in order to use the k-means clustering algorithm, the user's historical consumption data and product type need to be converted into feature vectors. It is worth noting that One-Hot encoding, TF-IDF (term frequency-inverse document frequency), and numerical normalization can be used to convert the user's historical consumption data and product type into feature vectors. It is worth noting that in order to avoid the influence of numerical features such as price on the clustering results, the features need to be normalized. Normalization adjusts all feature values ​​to the same numerical range (such as 0 to 1). Next, through the clustering algorithm, users with similar consumption behaviors are divided into several groups (clusters). The center point (center of mass) of each cluster represents the typical preference characteristics of the group.

[0079] Specifically, the clustering process includes: setting the number of clusters k, and randomly selecting k user feature vectors as the initial centroids. Calculating the distance from each user to the k centroids, and assigning the user to the cluster with the nearest centroid. For each cluster, calculate the average value of all user feature vectors in the cluster and update it as the new centroid. Repeat "cluster assignment" and "centroid update" until the centroid position no longer changes or the maximum number of iterations is reached. After clustering is completed, users are assigned to different groups, and each group has a corresponding center point that reflects the average preference characteristics of the group.

[0080] Specifically, the preference score reflects the user's interest in different types of goods. By comparing the distance between the user and the cluster center, the user's matching degree for each type of goods is calculated. In one achievable method, first, the distance from the user's feature value to the cluster center is calculated. For example, the Euclidean distance is used: Among them, x i is the user characteristic value, μ i is the characteristic value of the cluster center. Next, calculate the preference score, Where x is the user preference score. It is worth noting that the scores of all product types can be adjusted to a sum of 1 for comparison. For example, the scores of all product types are adjusted to within 0-1 using normalized calculation.

[0081] In step S13, a decision matrix is ​​constructed according to the user preference scores, and a weighted score matrix is ​​obtained by performing weighted calculation according to the user preference scores and the decision matrix.

[0082] Preferably, performing weighted calculation according to the user preference score and the decision matrix to obtain a weighted score matrix includes:

[0083] According to the user preference scores, a standardized score is calculated to obtain a standardized score matrix; wherein the standardized score matrix is ​​composed of the standardized scores;

[0084] According to the decision matrix, calculating the column average of each column in the decision matrix;

[0085] Calculate the column standard deviation based on the column average and the user preference score in each column;

[0086] According to the column standard deviation, weight calculation is performed to obtain column weight;

[0087] The column weight is multiplied by each element in the normalized score matrix to obtain a weighted score matrix.

[0088] Preferably, the step of calculating the standardized scores according to the user preference scores to obtain a standardized score matrix includes:

[0089] Calculating a standardized score based on the user preference score;

[0090] A standardized scoring matrix is ​​constructed according to the standardized scores; wherein the standardized scoring matrix is ​​composed of the standardized scores.

[0091] Preferably, calculating the standardized score according to the user preference score includes:

[0092] The normalized score is calculated using the following formula:

[0093]

[0094] In the formula, y ij is the standardized score, x ij Score user preference, min(x j ) represents the minimum user preference score of the jth column, max(x j ) represents the maximum user preference score of the jth column.

[0095] Preferably, calculating the column standard deviation according to the column average value and the user preference score in each column includes:

[0096] The column standard deviation is calculated by the following formula:

[0097]

[0098] In the formula, σ j is the column standard deviation, x ij Rate user preferences, is the column average and n is the total number of rows.

[0099] Preferably, performing weight calculation according to the column standard deviation to obtain the column weight includes:

[0100] The column weight is calculated by the following formula:

[0101]

[0102] In the formula, w j is the column weight, σ j is the column standard deviation, Represents the sum of the column standard deviations.

[0103] Specifically, in order to construct a decision matrix based on user preference scores and obtain a weighted score matrix through weighted calculation, it is necessary to follow the following steps in detail. First, a decision matrix is ​​constructed, in which the rows of the matrix represent alternatives (such as different commodities) and the columns represent the attributes of historical consumption commodities (such as style, brand, price). The elements in the matrix are directly filled with user preference scores, representing the user's preference for a certain attribute of an alternative. Next, a standardized score is calculated based on the user preference score to eliminate the influence of the dimensions and numerical ranges between different attributes, so as to facilitate subsequent comparisons. The standardized score is calculated by the formula, and all scores are converted to a range of 0 to 1 to form a standardized score matrix. Next, the column average of each column is first calculated based on the decision matrix. In one feasible method, the calculation formula for the column average is: Where n is the total number of rows. By calculating the average value of each column, the overall preference level of each attribute among all alternatives is obtained. Then, based on the column average value of each column and the elements in the decision matrix, the column standard deviation is calculated. The column standard deviation reflects the degree of dispersion of each attribute value. Next, the column weight is calculated based on the column standard deviation. The weight is used to reflect the importance of different attributes in the final score. The calculated weight can ensure that the weight value is between 0 and 1, and the sum of all weights is 1. Finally, the calculated column weight is multiplied by the standardized value of the user preference score to generate a weighted score matrix. Exemplarily, the calculation formula is: V ij =w j ·y ij , where w j is the element in the weighted scoring matrix, y ij Standardized scoring.

[0104] In step S14, the maximum value in the weighted scoring matrix is ​​selected to be determined as the positive ideal score, and the minimum value in the weighted scoring matrix is ​​selected to be determined as the negative ideal score.

[0105] It is worth noting that the positive ideal solution (PIS) represents the user's best preference for a certain attribute, and the maximum value of each column (corresponding to a certain attribute) in the weighted scoring matrix is ​​selected as the positive ideal score of the column. The negative ideal solution (NIS) represents the user's worst preference for a certain attribute, and the minimum value of each column in the weighted scoring matrix is ​​selected as the negative ideal score of the column.

[0106] Specifically, positive ideal solution (PIS) and negative ideal solution (NIS) belong to the TOPSIS method. It is worth noting that the TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution) method is a multi-attribute decision analysis (MADM) method for selecting the best solution or ranking the solutions among multiple alternatives. The core idea is to select the alternative that is close to the ideal solution and far away from the anti-ideal solution by comparing the distance between the alternative and the ideal solution (optimal solution) and the anti-ideal solution (worst solution). Specifically, in the present invention, a decision matrix is ​​constructed according to the user preference score, which is consistent with the first step in TOPSIS. Each row of the decision matrix represents an alternative solution, and each column represents a commodity attribute. The standardized score matrix is ​​calculated by standardized scoring, and a weighted score matrix is ​​generated in combination with the column weights, which is exactly the standardization and weighting step in the TOPSIS method. The maximum value of the weighted score matrix is ​​selected as the positive ideal score, and the minimum value is selected as the negative ideal score, which is consistent with the steps of determining the positive ideal solution and the negative ideal solution in the TOPSIS method.

[0107] In step S15, the distances between the positive ideal score, the negative ideal score and the standardized score are calculated respectively to obtain a positive ideal distance and a negative ideal distance; wherein the standardized score is calculated from the user preference score.

[0108] Specifically, the positive ideal score and the negative ideal score represent the user's best and worst evaluation of the product attributes, respectively. They are quantitative expressions of the user's ideal goals. By calculating the distance between each alternative and the positive ideal score and the negative ideal score, the pros and cons of each option can be quantified. The closer the distance is to the positive ideal score, the more it meets the user's needs; the farther the distance is from the negative ideal score, the less it meets the user's needs.

[0109] Specifically, the positive ideal distance represents the Euclidean distance between a solution and the positive ideal score. The positive ideal distance is calculated by the following formula: In the formula, is the positive ideal distance, is the value of the jth column in the positive ideal score vector. The negative ideal distance represents the Euclidean distance between a solution and the negative ideal score. The negative ideal distance is calculated using the following formula: In the formula, is the negative ideal distance, is the value of the jth column in the negative ideal score vector.

[0110] In step S16, the preference score proximity is calculated according to the positive ideal distance and the negative ideal distance, and the preference score proximity is sorted, and the alternative corresponding to the largest preference score proximity is selected as the recommendation result.

[0111] Preferably, the calculating the preference score proximity according to the positive ideal distance and the negative ideal distance comprises:

[0112] The preference score proximity is calculated by the following formula:

[0113]

[0114] In the formula, is the negative ideal distance, is the positive ideal distance, C i Score proximity for preferences.

[0115] Specifically, the preference score proximity C i It is an indicator used to measure the closeness of an alternative solution to the ideal solution, indicating the degree to which the solution is more in line with user preferences. i The closer it is to 1, the closer the solution is to the ideal solution and the more it meets user needs.

[0116] Specifically, the positive ideal distance and negative ideal distance of each solution are calculated in the previous step. Then, in an optional way, the preference score proximity C of all alternative solutions is calculated. i Sort in descending order, C i The solution with the largest value is the one closest to the user's preference and is recommended first. Select the solution with the closest preference score C from the sorted results. i The largest solution is recommended.

[0117] In summary, the present invention provides a personalized shopping guide method based on artificial intelligence, including: obtaining basic information of a user and attributes of historically consumed goods; wherein the basic information of the user includes the gender, age and height of the user, and the attributes of historically consumed goods include the style and brand of the goods; performing preference prediction according to the basic information of the user and the attributes of historically consumed goods to obtain a user preference score; constructing a decision matrix according to the user preference score, and performing weighted calculation according to the user preference score and the decision matrix to obtain a weighted score matrix; wherein the decision matrix is ​​composed of the user preference scores, the rows of the decision matrix represent alternatives, and the columns of the decision matrix represent the attributes of historically consumed goods; selecting the maximum value in the weighted score matrix to determine as a positive ideal score, and selecting the minimum value in the weighted score matrix to determine as a negative ideal score; respectively calculating the distances between the positive ideal score, the negative ideal score and the standardized score to obtain a positive ideal distance and a negative ideal distance; wherein the standardized score is calculated from the user preference score; calculating the preference score proximity according to the positive ideal distance and the negative ideal distance, and sorting the preference score proximity, and selecting the alternative corresponding to the largest preference score proximity as the recommendation result.

[0118] In the present invention, the method can obtain the basic information of the user (such as gender, age, height) and the attributes of historical consumption goods (such as the style and brand of the goods), use artificial intelligence algorithms (k-means clustering algorithm and Markov chain algorithm) to predict preferences, and combine the TOPSIS method to perform recommendation calculations. Specifically, the basic information of the user and the attributes of historical consumption goods are obtained, k-means clustering is performed based on the basic information of the user (such as gender, age, height), the correlation between the user and the goods is analyzed, and the type of goods suitable for the user is preliminarily determined. Combined with the historical consumption goods attributes of the user, the purchase behavior transfer probability is analyzed by the Markov chain algorithm, and the user's interest preference is further accurately predicted. A decision matrix containing user preference scores is constructed, and the column standard deviation is calculated by the standardized score matrix to obtain the weight distribution. The TOPSIS method is used to combine the ideal score and the negative ideal score, calculate the proximity of the preference score and sort it, and finally filter out the product recommendation results that best meet the user's needs. In summary, the method can focus on the personalized needs of users and provide more accurate product recommendations for users' specific preferences and purchase intentions.

[0119] Reference Figure 2 The second embodiment of the present invention provides a personalized shopping guide system based on artificial intelligence, including:

[0120] A data acquisition module is used to acquire basic user information and attributes of historically consumed goods; wherein the basic user information includes the user's gender, age and height, and the attributes of historically consumed goods include the style and brand of the goods;

[0121] A preference prediction module, used to make preference predictions based on the basic information of the user and the attributes of the historically consumed goods, and obtain a user preference score;

[0122] A matrix construction module, used to construct a decision matrix according to the user preference scores, and perform weighted calculation according to the user preference scores and the decision matrix to obtain a weighted score matrix; wherein the decision matrix is ​​composed of the user preference scores, the rows of the decision matrix represent alternatives, and the columns of the decision matrix represent the attributes of the historical consumption commodities;

[0123] A score determination module, configured to select the maximum value in the weighted score matrix to be determined as a positive ideal score, and select the minimum value in the weighted score matrix to be determined as a negative ideal score;

[0124] A distance calculation module, used to calculate the distance between the positive ideal score, the negative ideal score and the standardized score, respectively, to obtain a positive ideal distance and a negative ideal distance; wherein the standardized score is calculated from the user preference score;

[0125] The recommendation calculation module is used to calculate the preference score proximity according to the positive ideal distance and the negative ideal distance, sort the preference score proximity, and select the alternative corresponding to the largest preference score proximity as the recommendation result.

[0126] In an optional implementation, the preference prediction module is specifically used to:

[0127] According to the basic information of the user, commodity prediction is performed based on the k-means clustering algorithm and the Markov chain algorithm to obtain the commodity type;

[0128] According to the attributes of the historically consumed goods and the types of goods, preference prediction is performed based on the k-means clustering algorithm to obtain a user preference score.

[0129] In an optional implementation, the matrix construction module is specifically used for:

[0130] According to the user preference scores, a standardized score is calculated to obtain a standardized score matrix; wherein the standardized score matrix is ​​composed of the standardized scores;

[0131] According to the decision matrix, calculating the column average of each column in the decision matrix;

[0132] Calculate the column standard deviation based on the column average and the user preference score in each column;

[0133] According to the column standard deviation, weight calculation is performed to obtain column weight;

[0134] The column weight is multiplied by each element in the normalized score matrix to obtain a weighted score matrix.

[0135] In an optional implementation, the matrix construction module is specifically used for:

[0136] Calculating a standardized score based on the user preference score;

[0137] A standardized scoring matrix is ​​constructed according to the standardized scores; wherein the standardized scoring matrix is ​​composed of the standardized scores.

[0138] It should be noted that the personalized shopping guide system based on artificial intelligence provided in an embodiment of the present invention is used to execute all the process steps of the personalized shopping guide method based on artificial intelligence in the above embodiment. The working principles and beneficial effects of the two correspond one to one, so they will not be repeated here.

[0139] An embodiment of the present invention further provides an electronic device. The electronic device includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a personalized shopping guide program based on artificial intelligence. When the processor executes the computer program, the steps in the above-mentioned embodiments of the personalized shopping guide method based on artificial intelligence are implemented, such as Figure 1 Alternatively, when the processor executes the computer program, the functions of the modules / units in the above-mentioned device embodiments are implemented, such as the matrix construction module.

[0140] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program in the electronic device.

[0141] The electronic device may be a computing device such as a desktop computer, a notebook, a PDA, and a smart tablet. The electronic device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. The electronic device may include more or fewer components than the above components, or may combine certain components, or different components. For example, the electronic device may also include input and output devices, network access devices, buses, etc.

[0142] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the electronic device, and uses various interfaces and lines to connect various parts of the entire electronic device.

[0143] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the electronic device by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0144] Wherein, if the module / unit integrated in the electronic device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-0nly Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0145] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0146] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A personalized shopping guide method based on artificial intelligence, characterized in that: include: Obtaining basic user information and attributes of historically consumed goods; wherein the basic user information includes the user's gender, age, and height, and the attributes of historically consumed goods include the style and brand of the goods; According to the basic information of the user and the attributes of the historically consumed goods, preference prediction is performed to obtain a user preference score; A decision matrix is ​​constructed according to the user preference scores, and a weighted score matrix is ​​obtained by performing weighted calculation according to the user preference scores and the decision matrix; wherein the decision matrix is ​​composed of the user preference scores, the rows of the decision matrix represent alternatives, and the columns of the decision matrix represent the attributes of the historically consumed goods; Select the maximum value in the weighted scoring matrix to determine it as a positive ideal score, and select the minimum value in the weighted scoring matrix to determine it as a negative ideal score; Calculating the distances between the positive ideal score, the negative ideal score and the standardized score respectively to obtain a positive ideal distance and a negative ideal distance; wherein the standardized score is calculated from the user preference score; The preference score proximity is calculated according to the positive ideal distance and the negative ideal distance, and the preference score proximity is sorted, and the alternative corresponding to the largest preference score proximity is selected as the recommendation result.

2. The personalized shopping guide method based on artificial intelligence according to claim 1, characterized in that: The step of performing preference prediction based on the basic information of the user and the attributes of the historically consumed commodities to obtain a user preference score includes: According to the basic information of the user, commodity prediction is performed based on the k-means clustering algorithm and the Markov chain algorithm to obtain the commodity type; According to the attributes of the historically consumed goods and the types of goods, preference prediction is performed based on the k-means clustering algorithm to obtain a user preference score.

3. The personalized shopping guide method based on artificial intelligence according to claim 1, characterized in that: The weighted scoring matrix is ​​obtained by performing weighted calculation according to the user preference score and the decision matrix, including: According to the user preference scores, a standardized score is calculated to obtain a standardized score matrix; wherein the standardized score matrix is ​​composed of the standardized scores; According to the decision matrix, calculating the column average of each column in the decision matrix; Calculate the column standard deviation based on the column average and the user preference score in each column; According to the column standard deviation, weight calculation is performed to obtain column weight; The column weight is multiplied by each element in the normalized score matrix to obtain a weighted score matrix.

4. The personalized shopping guide method based on artificial intelligence according to claim 3 is characterized in that: The step of calculating the standardized scores according to the user preference scores to obtain a standardized score matrix includes: Calculating a standardized score based on the user preference score; A standardized scoring matrix is ​​constructed according to the standardized scores; wherein the standardized scoring matrix is ​​composed of the standardized scores.

5. The personalized shopping guide method based on artificial intelligence according to claim 4 is characterized in that: The step of calculating a standardized score according to the user preference score comprises: The normalized score is calculated using the following formula: In the formula, y ij is the standardized score, x ij Score user preference, min(x j ) represents the minimum user preference score of the jth column, max(x j ) represents the maximum user preference score of the jth column.

6. The personalized shopping guide method based on artificial intelligence according to claim 3, characterized in that: The calculating the column standard deviation according to the column average value and the user preference score in each column includes: The column standard deviation is calculated by the following formula: In the formula, σ j is the column standard deviation, x ij Rate user preferences, is the column average and n is the total number of rows.

7. The personalized shopping guide method based on artificial intelligence according to claim 3, characterized in that: The step of performing weight calculation according to the column standard deviation to obtain the column weight includes: The column weight is calculated by the following formula: In the formula, w j is the column weight, σ j is the column standard deviation, Represents the sum of the column standard deviations.

8. The personalized shopping guide method based on artificial intelligence according to claim 1, characterized in that: The calculating the preference score proximity according to the positive ideal distance and the negative ideal distance comprises: The preference score proximity is calculated by the following formula: In the formula, is the negative ideal distance, is the positive ideal distance, C i Score proximity for preferences.

9. A personalized shopping guide system based on artificial intelligence, characterized in that: include: A data acquisition module is used to acquire basic user information and attributes of historically consumed goods; wherein the basic user information includes the user's gender, age and height, and the attributes of historically consumed goods include the style and brand of the goods; A preference prediction module, used to make preference predictions based on the basic information of the user and the attributes of the historically consumed goods, and obtain a user preference score; A matrix construction module, used to construct a decision matrix according to the user preference scores, and perform weighted calculation according to the user preference scores and the decision matrix to obtain a weighted score matrix; wherein the decision matrix is ​​composed of the user preference scores, the rows of the decision matrix represent alternatives, and the columns of the decision matrix represent the attributes of the historical consumption commodities; A score determination module, configured to select the maximum value in the weighted score matrix to be determined as a positive ideal score, and select the minimum value in the weighted score matrix to be determined as a negative ideal score; A distance calculation module, used to calculate the distance between the positive ideal score, the negative ideal score and the standardized score, respectively, to obtain a positive ideal distance and a negative ideal distance; wherein the standardized score is calculated from the user preference score; The recommendation calculation module is used to calculate the preference score proximity according to the positive ideal distance and the negative ideal distance, sort the preference score proximity, and select the alternative corresponding to the largest preference score proximity as the recommendation result.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the personalized shopping guide method based on artificial intelligence as described in any one of claims 1 to 8.

Images