A motor product intelligent recommendation method based on similarity evaluation

Abstract

The application provides a motor product intelligent recommendation method based on similarity evaluation, comprising: obtaining motor product performance parameter data input by a user, extracting input parameter mapping to a vector space to obtain a first feature vector; calling product performance parameter data matched with the input parameter pre-stored in a product database, mapping the product performance parameter data to the vector space to obtain a plurality of second feature vectors; uniformly performing normalization processing on parameters in the first feature vector and the plurality of second feature vectors; calculating the similarity between the first feature vector and the plurality of second feature vectors by using an improved Euclidean distance method, and sorting the products according to the product similarity from large to small to generate a motor product recommendation list for the customer. By the application, the accuracy of product recommendation and the speed of customer selection are improved, and the personalized needs of the customer are effectively met.

Description

Technical Field

[0001] This invention relates to the field of industrial production, specifically to an intelligent recommendation method for motor products based on similarity evaluation. Background Technology

[0002] In the industrial production sector, the current diversification and complexity of product selection have made effective product selection difficult. Take Kollmorgen as an example: their product selection manual is 45 pages long, containing a large amount of images, text, and product parameters. This is disadvantageous for customers looking to choose the right product. Successful selection requires not only extensive professional knowledge but also a good understanding of the manual. Even if these two conditions are met, customers still need to do considerable reading and comparison to find the product they need. Furthermore, when products are updated, the selection manual is not updated in a timely manner, preventing customers from accessing the latest products. These factors not only slow down the customer's selection process but also negatively impact their experience.

[0003] Based on previous research on similarity algorithms and combined with relevant knowledge in the field of industrial production, this invention proposes an intelligent recommendation method for motor products based on similarity evaluation to effectively solve the above-mentioned problems. Summary of the Invention

[0004] To address the shortcomings of existing methods, this invention provides an intelligent recommendation method for motor products based on similarity evaluation, thereby solving the problem of customer selection difficulties caused by too many product options in the prior art.

[0005] This invention is based on a similarity evaluation intelligent algorithm. It uses an improved Euclidean distance method to calculate the similarity between the motor products required by the customer and the existing motor products in the product database. The products are sorted in descending order of similarity to generate a recommended list of motor products for the customer. This improves the accuracy of product recommendations and the speed of customer selection, effectively meeting the personalized needs of customers.

[0006] This invention provides an intelligent recommendation method for motor products based on similarity evaluation, comprising:

[0007] S1: Obtain the motor product performance parameter data input by the user. The motor product performance parameter data includes rated power, speed, torque, and voltage. Fill in the empty positions in the input parameters with 0. Extract the user's input parameters and map them to a vector space to obtain a first feature vector. Retrieve the product performance parameter data that matches the input parameters from the product database. Make the corresponding product performance parameters that match the input parameters that are filled in with 0 also fill in with 0. Map the product performance parameter data to a vector space to obtain several second feature vectors.

[0008] S2: The parameters in the first feature vector and the plurality of second feature vectors are uniformly normalized. The formula for the normalization process is as follows:

[0009]

[0010] Where i represents the serial number of the motor product already existing in the product database, and k represents the serial number of the performance parameter of the i-th motor product. s represents the average of the corresponding performance parameters. k x represents the variance of the corresponding performance parameter. ik x represents the k-th performance parameter of the i-th motor product. , ik The standardized value of the k-th performance parameter of an i-th motor product;

[0011] S3: Calculate the similarity between the first feature vector and the plurality of second feature vectors using an improved Euclidean distance method. This improved Euclidean distance method primarily uses Euclidean distance and secondarily uses cosine similarity. The products are sorted from highest to lowest similarity to generate a recommended list of motor products for the customer. Specifically, step S3 involves:

[0012] S31: Calculate the Euclidean distance between the first feature vector and the plurality of second feature vectors using the Euclidean distance algorithm, sorting them in ascending order. The formula for the Euclidean distance algorithm is as follows:

[0013]

[0014] Where α represents the vector of input motor product parameters, β i The vector represents the number of the i-th motor product parameters already existing in the product database, where j represents the motor product performance parameter number, and α j β represents the j-th performance parameter vector representing the input motor product parameters. ij This represents the j-th performance parameter vector of the i-th motor product already existing in the product database, dist(α,β). i Both ) and D represent the Euclidean distance between the two vectors;

[0015] S32: When the Euclidean distance between the first feature vector and the plurality of second feature vectors is the same, the cosine values ​​of the first feature vector and the plurality of second feature vectors with the same Euclidean distance are calculated according to the cosine value formula, and sorted in descending order of cosine value. The cosine value calculation formula is as follows:

[0016]

[0017] Where α represents a vector of input motor product parameters; β iThe vector representing the parameters of the i-th motor product already existing in the product database; α j The vector representing the j-th performance parameter of the input motor product parameters; β j represents the j-th performance parameter vector of an existing motor product in the product database; k represents the motor product performance parameter number; C represents the cosine value between the two vectors.

[0018] S33: Calculate the similarity of motor products using a formula, sort the products from highest to lowest similarity, and generate a recommended list of motor products for the customer. The formula is:

[0019]

[0020] Where α represents a vector of input motor product parameters; β i The vector representing the parameters of the i-th motor product already existing in the product database; dist(α,β) i ) represents the Euclidean distance between two vectors; S represents the similarity between the two vectors.

[0021] Preferably, when in step S31, if two or more product performance parameters preset in the product database have the same Euclidean distance as the input parameter, a cosine similarity method is further used to assist in the judgment, and the products with larger cosine values ​​are ranked first and recommended first.

[0022] Preferably, after generating the recommended list of motor products, when a customer selects a recommended motor, the recommended motor model is stored, and the customer can click Next to repeat steps S1-S3 to recommend associated encoders. The encoder performance parameter data includes type, outer diameter, resolution, and voltage. When no recommended motor is selected, the input parameters are stored, and the recommendation process ends.

[0023] Preferably, after generating the recommended list of motor products, when a customer selects a recommended motor, the recommended motor model is stored, and the customer can click Next to repeat steps S1-S3 to recommend associated servo controllers. The servo controller performance parameter data includes power and voltage. When no recommended motor is selected, the input parameters are stored, and the recommendation process ends.

[0024] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0025] 1. This invention can be effectively applied to the industrial production field, making it easier for enterprises to efficiently recommend motor products to customers;

[0026] 2. This invention is based on a similarity evaluation intelligent algorithm. It uses an improved Euclidean distance method to calculate the similarity between the motor products required by the customer and the existing motor products in the product database. The products are sorted in descending order of similarity to generate a recommended list of motor products for the customer. This improves the accuracy of product recommendations and the speed of customer selection, effectively meeting the personalized needs of customers. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is a flowchart of the intelligent recommendation method for motor products based on similarity evaluation in an embodiment of the present invention. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] In the description of this invention, it should be noted that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0031] A method for intelligent recommendation of motor products based on similarity evaluation includes the following steps:

[0032] S1: Obtain the motor product performance parameter data input by the user. The motor product performance parameter data includes rated power, speed, torque, voltage, wiring method, and cooling method. Fill in the empty positions in the input parameters with 0. Extract the user's input parameters and map them to a vector space to obtain a first feature vector. Retrieve the product performance parameter data that matches the input parameters from the product database. Make the corresponding product performance parameters that match the input parameters that are filled in with 0 also filled in with 0. Map the product performance parameter data to a vector space to obtain several second feature vectors.

[0033] Because there are many types and parameters of motor products, and many factors influence the selection, it is necessary to first extract the performance parameter data of the motor products so that customers can make personalized choices.

[0034] Meanwhile, this invention does not limit the method of mapping the input parameters and the product performance parameters to a vector space.

[0035] S2: Normalize the parameters in the first feature vector and the plurality of second feature vectors.

[0036] Preferably, since different units of measurement have a significant impact on the calculation results, the parameters in the first feature vector and the plurality of second feature vectors in step S2 are normalized. The formula for the normalization process is as follows:

[0037]

[0038] Where i represents the serial number of the motor product already existing in the product database, and k represents the serial number of the performance parameter of the i-th motor product. s represents the average of the corresponding performance parameters. k x represents the variance of the corresponding performance parameter. ik x represents the k-th performance parameter of the i-th motor product. , ik The standardized value of the k-th performance parameter of an i-th motor product.

[0039] Different units of measurement can lead to measurement errors and affect the calculation results. Therefore, by normalizing each parameter using this formula, the influence caused by different units of measurement can be eliminated, and the similarity of the calculation can be improved.

[0040] S3: Calculate the similarity between the first feature vector and the plurality of second feature vectors using an improved Euclidean distance method. The improved Euclidean distance method is based on Euclidean distance and supplemented by cosine similarity. Sort the products according to their similarity from largest to smallest to generate a recommended list of motor products for the customer.

[0041] Preferably, when producing a recommendation list for customers, if a customer selects a motor product based on the three performance parameters of rated power, speed, and torque, the products in the product database are also compared with the products required by the customer based solely on the three performance parameters of rated power, speed, and torque.

[0042] Euclidean distance refers to the true distance between two points in m-dimensional space. A larger distance indicates lower similarity, and vice versa. In recommending motor products, the similarity we use is primarily for calculating the similarity between two vectors, which is essentially finding the distance between n-dimensional vectors. Therefore, the expression for n-dimensional space is as follows:

[0043]

[0044] Where x i y i Represents a vector; n represents the dimension; ρ d＝n This represents the Euclidean distance between two vectors.

[0045] While Euclidean distance can effectively calculate products that closely match customer needs, it is typically used to calculate the similarity between two vectors. This invention aims to calculate the Euclidean distance between the input parameter and multiple parameters pre-set in a product database. During the calculation, two or more product performance parameters may have the same Euclidean distance as the input parameter, affecting the accuracy of the recommendation results. Therefore, this invention utilizes an improved Euclidean distance method to calculate the similarity between the input parameter and the multiple parameters pre-set in the product database. This improved Euclidean distance method primarily uses Euclidean distance, supplemented by cosine similarity. When two or more product performance parameters have the same Euclidean distance as the input parameter, cosine similarity is used to calculate the cosine value of the user parameter with the same Euclidean distance and the parameters pre-set in the product database. This compensates for the shortcomings of Euclidean distance calculation and further optimizes the accuracy of the recommendation results.

[0046] Preferably, step S3 specifically comprises:

[0047] S31: Calculate the Euclidean distance between the first feature vector and the plurality of second feature vectors using the Euclidean distance algorithm, sorting them in ascending order. The formula for the Euclidean distance algorithm is as follows:

[0048]

[0049] Where α represents the vector of input motor product parameters, β i The vector represents the number of the i-th motor product parameters already existing in the product database, where j represents the motor product performance parameter number, and α j β represents the j-th performance parameter vector representing the input motor product parameters. ij This represents the j-th performance parameter vector of the i-th motor product already existing in the product database, dist(α,β). i Both ) and D represent the Euclidean distance between the two vectors;

[0050] S32: When the Euclidean distance between the first feature vector and the plurality of second feature vectors is the same, the cosine values ​​of the first feature vector and the plurality of second feature vectors with the same Euclidean distance are calculated according to the cosine value formula, and sorted in descending order of cosine value. The cosine value calculation formula is as follows:

[0051]

[0052] Where α represents a vector of input motor product parameters; β i The vector representing the parameters of the i-th motor product already existing in the product database; α j The vector representing the j-th performance parameter of the input motor product parameters; β j represents the j-th performance parameter vector of an existing motor product in the product database; k represents the motor product performance parameter number; C represents the cosine value between the two vectors.

[0053] Cosine similarity measures the angle between vectors i and j. The smaller the angle, the larger the cosine value, indicating that the two products are more similar and thus have a higher recommendation priority.

[0054] Preferably, when in step S31, if two or more product performance parameters preset in the product database have the same Euclidean distance as the input parameter, a cosine similarity method is further used to assist in the judgment, and the products with larger cosine values ​​are ranked first and recommended first.

[0055] S33: Calculate the similarity of motor products using a formula, sort the products from highest to lowest similarity, and generate a recommended list of motor products for the customer. The formula is:

[0056]

[0057] Where α represents a vector of input motor product parameters; β i The vector representing the parameters of the i-th motor product already existing in the product database; dist(α,β) i ) represents the Euclidean distance between two vectors; S represents the similarity between the two vectors.

[0058] Preferably, after generating the recommended list of motor products, when a customer selects a recommended motor, the recommended motor model is stored, and the customer can click Next to repeat steps S1-S3 to recommend associated encoders. The encoder performance parameter data includes type, outer diameter, resolution, and voltage. When no recommended motor is selected, the input parameters are stored, and the recommendation process ends.

[0059] Preferably, after generating the recommended list of motor products, when a customer selects a recommended motor, the recommended motor model is stored, and the customer can click Next to repeat steps S1-S3 to recommend associated servo controllers. The servo controller performance parameter data includes power and voltage. When no recommended motor is selected, the input parameters are stored, and the recommendation process ends.

[0060] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent recommendation of motor products based on similarity evaluation, characterized in that, include: S1: Obtain the motor product performance parameter data input by the user. The motor product performance parameter data includes rated power, speed, torque, and voltage. Fill in the empty positions in the input parameters with 0. Extract the user's input parameters and map them to the vector space to obtain the first feature vector. Retrieve product performance parameter data that matches the input parameter from the product database, and make the corresponding product performance parameter that matches the input parameter that is occupied by 0 also occupy by 0. Map the product performance parameter data to the vector space to obtain several second feature vectors. S2: The parameters in the first feature vector and the plurality of second feature vectors are uniformly normalized. The formula for the normalization process is as follows: , Where i represents the serial number of the motor product already existing in the product database, and k represents the serial number of the performance parameter of the i-th motor product. This represents the average of the corresponding performance parameters. This represents the variance of the corresponding performance parameter. This represents the k-th performance parameter of the i-th motor product. The standardized value of the k-th performance parameter of an i-th motor product; S3: Calculate the similarity between the first feature vector and the plurality of second feature vectors using an improved Euclidean distance method. This improved Euclidean distance method primarily uses Euclidean distance and secondarily uses cosine similarity. The products are sorted from highest to lowest similarity to generate a recommended list of motor products for the customer. Specifically, step S3 involves: S31: Calculate the Euclidean distance between the first feature vector and the plurality of second feature vectors using the Euclidean distance algorithm, sorting them in ascending order. The formula for the Euclidean distance algorithm is as follows: , in, A vector representing the input motor product parameters. This represents the vector containing the parameters of the i-th motor product in the product database, where j represents the serial number of the motor product's performance parameters. This represents the j-th performance parameter vector representing the input motor product parameters. This means that the product database already contains the j-th performance parameter vector for the i-th motor product. and Both represent the Euclidean distance between two vectors; S32: When the Euclidean distance between the first feature vector and the plurality of second feature vectors is the same, the cosine values ​​of the first feature vector and the plurality of second feature vectors with the same Euclidean distance are calculated according to the cosine value formula, and sorted in descending order of cosine value. The cosine value calculation formula is as follows: , in, A vector representing the input motor product parameters; This represents a vector containing the parameters of the i-th motor product that is already present in the product database. The vector representing the j-th performance parameter of the input motor product parameters; represents the j-th performance parameter vector of an existing motor product in the product database; k represents the motor product performance parameter number; C represents the cosine value between the two vectors. S33: Calculate the similarity of motor products using a formula, sort the products from highest to lowest similarity, and generate a recommended list of motor products for the customer. The formula is: , in, A vector representing the input motor product parameters; This represents a vector containing the parameters of the i-th motor product that is already present in the product database. represents the Euclidean distance between two vectors; S represents the similarity between the two vectors. In addition, when two or more product performance parameters preset in the product database have the same Euclidean distance as the input parameter in step S31, the cosine similarity method is further used to assist in the judgment, and the products with larger cosine values ​​are ranked first and recommended first. In addition, after generating the recommended motor product list, when a customer selects a recommended motor, the recommended motor model is stored. The customer can then click "Next" to repeat steps S1-S3 to recommend associated encoders. If no recommended motor is selected, the input parameters are stored, and the recommendation process ends. The encoder performance parameters include type, outer diameter, resolution, and voltage. Similarly, after generating the recommended motor product list, when a customer selects a recommended motor, the recommended motor model is stored. The customer can then click "Next" to repeat steps S1-S3 to recommend associated servo controllers. If no recommended motor is selected, the input parameters are stored, and the recommendation process ends. The servo controller performance parameters include power and voltage.

Images