Case reasoning-based regional food short video recommendation method and device

Abstract

The invention relates to a case reasoning-based regional food short video recommendation method and device. Compared with the prior art, the scheme mainly depends on keywords and user tags to search target cases; according to the method, on the basis of the case reasoning technology theory, the similarity of the videos and the cases in the aspect of recommendation attributes is firstly calculated, then the target case is finally calculated in combination with the weights of the recommendation attributes, user interests and the recommendation value of the cases can be better reflected, and the recommendation accuracy is improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of short video recommendation, and in particular, to a method and device for recommending regional food short videos based on case-based reasoning.

[0002] Background Art With the rapid development of short video platforms, recommending food culture through short videos has become an important way to understand traditional cuisine and food culture. As a new media form, short videos have the characteristics of fast information dissemination and rich and diverse content, and can vividly and intuitively display the production process, taste characteristics and cultural background of local foods, attracting the attention and love of a large number of users.

[0003] However, existing short video recommendation methods mainly rely on simple similarity calculations or user interest tag matching, and there are still some deficiencies in terms of recommendation accuracy and user experience. Specifically, the existing recommendation methods mainly include the following two types: First, one method is to extract the keywords of the videos that the user is interested in, and then calculate the similarity based on these keywords to find the videos to be recommended with similar similarity for recommendation. Although this method can capture the user's interests to a certain extent, it often relies too much on text information and ignores other important features of the video content, such as visual effects, narrative structures, etc.

[0004] Second, another method is to define the user's interest factors (or interest tags), and find the videos to be recommended that match based on these factors or tags. This method can better reflect the user's interest preferences, but may lead to overly single recommendation results and it is difficult to provide users with a diverse content experience.

[0005] The common defect of these existing methods is that they ignore some important recommendation attributes of short videos themselves. For example, for food short videos, in addition to the content theme, factors such as the shooting skills, color style, narrative structure, and functional value (such as tutorial type, store exploration type, cultural type) of the video may all affect the user's viewing experience and interest level. Therefore, there is a defect of low recommendation accuracy. Summary of the Invention

[0006] The following is an overview of the subject matter described in detail in this article. This overview is not intended to limit the scope of protection of the claims.

[0007] The main purpose of the embodiments of the present disclosure is to propose a method and device for recommending regional food short videos based on case-based reasoning, which can improve the recommendation accuracy of food short videos.

[0008] The first aspect of the embodiments of the present application proposes a method for recommending regional food short videos based on case-based reasoning, and the method includes: Obtain a target video and a case library, where there are M cases in the case library, and N first attribute values corresponding to each case on N recommended attributes; the cases are short food videos of the target area; the target video is the currently viewed short video; N and M are positive integers greater than 1; Determine N second attribute values corresponding to the target video on the N recommended attributes; According to the N first attribute values and the N second attribute values, calculate N first similarities between the target video and any case on the N recommended attributes; According to the N first similarities and the weights of the N recommended attributes, calculate a second similarity between the target video and any case; Determine a target case from the M cases according to the second similarity; Push the short food video corresponding to the target case.

[0009] A method for recommending short food videos of a region based on case-based reasoning provided by an embodiment of the present disclosure has at least the following beneficial effects: This method constructs N recommended attributes corresponding to each case and determines the corresponding N first attribute values, and uses the N first attribute values to represent the recommendation value of the case on the corresponding recommended attribute; then determines the N second attribute values of the target video; based on the first attribute values and the second attribute values, the similarity between the target video and any case on each recommended attribute can be determined; then the weights of the recommended attributes are introduced, and the global similarity between the target video and the case is calculated based on the weights and the similarities, and finally the target case is searched and recommended based on the global similarity.

[0010] Compared with the prior art solution that mainly relies on keywords and user tags to find target cases; this method is based on the theory of case-based reasoning technology, first calculates the similarity between the video and the case on the recommended attributes, and then combines the weights of the recommended attributes to finally calculate the target case, which can better reflect the user's interests and the recommendation value of the case itself, and improve the accuracy of the recommendation.

[0011] In some embodiments, before calculating the second similarity between the target video and any case according to the N first similarities and the weights of the N recommended attributes, it further includes: Calculate the weight of each recommended attribute:

[0012]

[0013]

[0014] Wherein, is the weight of the th recommended attribute, is the normalized value of the th case on the th recommended attribute, is the minimum value of the th case on the th recommended attribute, is the maximum value of the th case on the th recommended attribute, is the first attribute value of the th case on the th recommended attribute; is the information entropy of the th recommended attribute, is the natural logarithm function; The process of calculating the second similarity between the target video and any of the cases includes:

[0015] Wherein, is the target video, is the target video and the th case on the th recommended attribute, is the first similarity between the target video and the th case is the second similarity.

[0016] In some embodiments, after pushing the food short video corresponding to the target case, it further includes: Judging the degree of interest in the target case; If the degree of interest is less than the threshold, then select n recommended attributes from the N recommended attributes in descending order of weight; n is less than N and n is a positive integer; Determine the n second attribute values corresponding to the target video on the n recommended attributes, and determine the n second attribute values corresponding to the target video on the n recommended attributes; Calculate the n first similarities between the target video and any of the cases on the n recommended attributes according to the n first attribute values and the n second attribute values; Calculate the third similarity between the target video and any of the cases according to the n first similarities and the weights of the n recommended attributes; Determine the target case again from the M cases according to the third similarity and push it.

[0017] In some embodiments, the determining the target case from the M cases according to the second similarity includes: Based on the second similarity, sort the M cases in descending order to obtain a descending queue; Select the first case in the descending queue as the target case.

[0018] In some embodiments, the determining method of the degree of interest includes the following steps: Confirm the viewing duration of the target case; Calculate the ratio between the viewing duration and the total duration of the target case as the degree of interest.

[0019] In some embodiments, the recommended attributes include: core content attribute, visual presentation attribute, narrative structure attribute, functional value attribute, communication value attribute; The core content attribute includes: local cuisine, street snacks, creative cuisine; The visual presentation attribute includes: shot language, color style; The narrative structure attribute includes: golden 5-second rule, information density; The functional value attribute includes: tutorial type, store visit type, cultural type; The communication value attribute includes: view count, like count.

[0020] In some embodiments, before obtaining the target video and the case library, the method further includes: Obtain multi-modal food data of the target area; Construct multiple food short videos based on the multi-modal food data; Build the case library based on the multiple food short videos.

[0021] A second aspect of the embodiments of the present application proposes a regional food short video recommendation device based on case-based reasoning, and the device includes: A data acquisition module, configured to acquire a target video and a case library, where there are M cases in the case library, and N first attribute values corresponding to each case on N recommended attributes; the case is a food short video of the target area; the target video is the currently viewed short video; N and M are positive integers greater than 1; An attribute value determination module, configured to determine N second attribute values corresponding to the target video on the N recommended attributes; The first similarity calculation module is configured to calculate N first similarities between the target video and any of the cases on N recommended attributes according to the N first attribute values and the N second attribute values. The second similarity calculation module is configured to calculate the second similarity between the target video and any of the cases according to the N first similarities and the weights of the N recommended attributes. The case selection module is configured to determine a target case from M cases according to the second similarity. The case push module is configured to push the food short video corresponding to the target case.

[0022] In a third aspect of the embodiments of the present application, an electronic device is provided, including at least one controller and a memory communicatively connected to the at least one controller; the memory stores instructions executable by the at least one controller, and when the instructions are executed by the controller, the controller is caused to execute the case-based reasoning-based regional food short video recommendation method as described in the first aspect.

[0023] In a fourth aspect of the embodiments of the present application, a computer-readable storage medium is provided, where the computer-readable storage medium stores computer-executable instructions for causing a computer to execute the case-based reasoning-based regional food short video recommendation method as described above.

[0024] It can be understood that the beneficial effects of the above second aspect to the fourth aspect compared with the related art are the same as those of the first aspect compared with the related art, and reference may be made to the related descriptions in the first aspect, which will not be elaborated herein. Description of the Drawings

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for use in the embodiments or the related art. Obviously, the following drawings are only some embodiments of the present application, and for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0026] Figure 1 is a flowchart of the case-based reasoning-based regional food short video recommendation method provided by the embodiments of the present application; Figure 2 is a flowchart of case-based reasoning; Figure 3 is a flowchart of the case-based reasoning-based regional food short video recommendation method provided by another embodiment of the present application; Figure 4 is a structural diagram of the case-based reasoning-based regional food short video recommendation device provided by the embodiments of the present application; Figure 5 It is a schematic diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0027] In order to make the objectives, technical solutions and advantages of the present application more clear and understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0028] It should be noted that although functional module division is performed in the device schematic diagram and the logical sequence is shown in the flowchart, in some cases, the steps shown or described can be executed in a different module division in the device or a different sequence in the flowchart. Terms such as "first" and "second" in the description, claims and the above-mentioned drawings are used to distinguish similar objects and do not necessarily need to describe a specific order or sequence.

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application. With the rapid development of short video platforms, recommending food culture through short videos has become an important way to understand traditional cuisine and food culture. As an emerging media form, short videos have the characteristics of fast information dissemination and rich and diverse content, and can vividly and intuitively display the production process, taste characteristics and cultural background of local foods, attracting the attention and love of a large number of users.

[0030] However, existing short video recommendation methods mainly rely on simple similarity calculations or user interest tag matching, and there are still some deficiencies in terms of recommendation accuracy and user experience. Specifically, the existing recommendation methods mainly include the following two types: First, one method is to extract the keywords of the videos that the user is interested in, and then calculate the similarity based on these keywords to find the videos to be recommended with similar similarity for recommendation. Although this method can capture the user's interests to a certain extent, it often relies too much on text information and ignores other important features of the video content, such as visual effects and narrative structures.

[0031] Second, another method is to define the user's interest factors (or interest tags), and find the matching videos to be recommended based on these factors or tags. This method can better reflect the user's interest preferences, but may lead to overly single recommendation results and it is difficult to provide users with a diverse content experience.

[0032] The common defect of these existing methods is that they ignore some important recommendation attributes of short videos themselves. For example, for food short videos, in addition to the content theme, factors such as shooting skills, color style, narrative structure, and functional value (such as tutorial type, store exploration type, cultural type) of the video may all affect the user's viewing experience and level of interest. Therefore, there is a defect of low recommendation accuracy.

[0033] As Figure 1 shown, an embodiment of the present application provides a method for recommending regional food short videos based on case-based reasoning. The method includes: Step S110, obtain a target video and a case base, where there are M cases in the case base.

[0034] Step S120, determine N second attribute values corresponding to the target video on N recommendation attributes.

[0035] Step S130, calculate N first similarity degrees between the target video and any case on N recommendation attributes according to the N first attribute values and the N second attribute values.

[0036] Step S140, calculate the second similarity degree between the target video and any case according to the N first similarity degrees and the weights of the N recommendation attributes.

[0037] Step S150, determine a target case from the M cases according to the second similarity degree.

[0038] Step S160, push the food short video corresponding to the target case.

[0039] In step S110, the target video refers to the food short video that the user watches on the client. There are M cases in the case base. Let the case base : .

[0040] The case is a food short video of the target area, and the target area is a specified area, such as Zhangjiajie, Wuling Mountain Area, etc.

[0041] In order to evaluate the value of the food short video itself and achieve the purpose of recommending the food in the target area, the present application introduces recommendation attributes. The recommendation attributes of the food short video can represent the characteristic information that conforms to specific attributes of the food short video. The recommendation attributes have corresponding recommendation attribute values, and the recommendation attribute values can represent the amount of information of the characteristic information.

[0042] There are M cases in the case base, and the food short video has N recommendation attributes. Let the recommendation attribute : .

[0043] Each case has a corresponding first attribute value for each recommended attribute as : .

[0044] In some embodiments, for example, the recommended attributes include: core content attributes, visual presentation attributes, narrative structure attributes, functional value attributes, and communication value attributes; 1. The core content attributes include: (1) Local cuisine, highlighting local characteristics (such as the spicy visual of Hunan cuisine and the delicate presentation of some dishes).

[0045] (2) Street snacks, with a sense of urban life (night markets, morning markets, mobile vendors).

[0046] (3) Creative cuisine, content for the curious (molecular cuisine, cross - border fusion dishes).

[0047] 2. The visual presentation attributes include: (1) Camera language, dynamic camera movement, macro lenses (bubbles, textures, sauce flow), contrast editing; (2) Color style, warm tones (red / yellow enhance appetite (frequently used in hot pot and barbecue)), high contrast (dark background + bright food colors (common in Western cuisine and desserts)), vintage film (a sense of nostalgia (for traditional snacks and childhood snacks themes)); 3. The narrative structure attributes include: (1) The golden 5 - second rule, with a highlight in the first 3 seconds (strong impact images such as pulled cheese, flame grilling, food splashing), the middle section rhythm (showing key steps (marinating → frying → plating) within 20 seconds), and an ending guide ("Tap the screen to get the same style" "Wait for the tutorial in the comment section"); (2) Information density, short duration (15 - 30 seconds, for example, focusing on the highlights of a single dish (such as "Finish sucking the crab noodles in 10 seconds")), long duration (1 minute +, adding a story - line of a food - tasting vlog or family food heritage); 4. The functional value attributes include: Tutorial type, teaching of home - cooked dishes, replication of internet - famous dishes; Food - tasting type, restaurant reviews, discovery of hidden menus; Cultural type, intangible cultural heritage food techniques, tracing the origin of food culture; 5. The communication value attributes include: Page views, likes.

[0048] It should be noted that the recommended attributes are set based on prior experience (such as expert systems). Although the recommended attributes of this application need to be set based on experience, in the recommendation process of all food short videos, a set of standard recommended attributes (that is, the categories of recommended attributes are the same, and the assignment criteria of recommended attributes are the same) are used for the judgment process; it is not because different food short videos use different judgment criteria for recommended attributes that recommendation errors occur. It should also be noted that the setting of the recommended attributes here is not the focus of this application. The focus lies in the subsequent process of recommending target cases based on case-based reasoning.

[0049] It should be noted that in the process of assigning values to the recommended attributes, the assignment criteria for each recommended attribute can be preset. For example, for the "communication value attribute", each view is assigned 1, and each like is assigned 5. If a food short video has 1000 views and 20 likes, then the corresponding value of the "communication value attribute" is 1100. Subsequently, in the calculation process of the similarity based on the recommended attribute values and the calculation process of the weights, the recommended attribute values of each recommended attribute will be normalized.

[0050] For step S120, determine the N second attribute values corresponding to the target video on N recommended attributes.

[0051] The judgment criteria for the target video are the same as those for the cases.

[0052] In step S130, according to the N first attribute values and the N second attribute values, calculate the N first similarities between the target video and any case on N recommended attributes.

[0053] The attribute values corresponding to the recommended attributes can be exact values (for example, the attribute value of the "communication value attribute" can be 1100), or they can be intervals or fuzzy vectors.

[0054] For the recommended attributes with exact attribute values, normalization and Euclidean distance can be used to calculate the first similarity between the target video and any case on a certain recommended attribute.

[0055] For the recommended attributes with interval attribute values, the Jaccard coefficient can be used for calculation. For the recommended attributes with fuzzy vector attribute values, the Wasserstein distance can be used for calculation and design.

[0056] In step S140, according to the N first similarities and the weights of the N recommended attributes, calculate the second similarity between the target video and any case. Among them, because there are multiple recommended attributes, different weights are designed for different recommended attributes to balance the importance of different recommended attributes. The following introduces how to obtain the weights: In this application, before calculating the second similarity between the target video and any case according to the N first similarities and the weights of the N recommended attributes in step S140, the method further includes: Step S210, calculating the weight of each recommended attribute: (1) (2) (3) Among them, is the weight of the th recommended attribute, is the normalization value of the th case on the th recommended attribute, is the minimum value of the th case on the th recommended attribute, is the maximum value of the th case on the th recommended attribute, is the first attribute value of the th case on the th recommended attribute; is the information entropy of the th recommended attribute, is the natural logarithm function.

[0057] Step S220, the process of calculating the second similarity between the target video and any case includes: (4) Among them, is the target video, is the first similarity between the target video and the th case on the th recommended attribute, is the second similarity between the target video and the

[0058] th case. Formula (1) normalizes each case on any one attribute, making its value range [0, 1].

[0059] Formula (2) calculates the information entropy corresponding to any recommended attribute. Since the data variation degree of the recommended attribute index is relatively large, the role of this recommended attribute in distinguishing different cases is more prominent. Therefore, a greater weight should be assigned. Vice versa. However, due to the poor discrimination of the recommended attributes with small weights, this paper uses the concept of the index variation degree in information entropy to characterize the discrete degree of the values between recommended attributes.

[0060] Formula (3) is the process of calculating weights based on information entropy. Information entropy (proposed by Shannon) is an index to measure the uncertainty or chaos degree of information. This application can use information entropy to characterize the discrete degree of the values between recommended attributes, which can improve the accuracy between the calculated weights.

[0061] Formula (4) is the process of calculating the second similarity based on the first similarity and weights.

[0062] Finally, according to the second similarity, the target case is determined from M cases, and the push is performed based on the target case.

[0063] The method provided by this application has at least the following beneficial effects: This method constructs N recommended attributes corresponding to each case and determines the corresponding N first attribute values, and uses the N first attribute values to characterize the recommendation value of the case on the corresponding recommended attribute; then determines the N second attribute values of the target video; based on the first attribute value and the second attribute value, the similarity between the target video and any case on each recommended attribute can be determined; then the weight of the recommended attribute is introduced, and the global similarity between the target video and the case is calculated based on the weight and the similarity. Finally, the target case is searched and recommended based on the global similarity.

[0064] Compared with the prior art that mainly relies on keywords and user tags to find the target case, ignoring the recommended attributes of the video itself, resulting in low recommendation accuracy; this method is based on the case-based reasoning technology theory, first calculates the similarity between the video and the case on the recommended attributes, and then combines the weights of the recommended attributes, and finally calculates the target case, which can better reflect the user's interest and the recommendation value of the case itself, and improve the recommendation accuracy.

[0065] Compared with the technical solution of dividing corresponding tags for each case, this method designs N recommended attributes, covering more recommendation directions. Each case has corresponding recommended attribute values on the N recommended attributes, which also reflects the recommendation value of the case on different recommended attributes, not limited to corresponding to a certain type of tag, so as to improve the accuracy.

[0066] As Figure 2 shown, after step S150, this method further includes: Step S310, judging the degree of interest in the target case; Step S320, if the degree of interest is less than the threshold, then select n recommended attributes from the N recommended attributes in the order of weight from largest to smallest; n is less than N, and n is a positive integer; Step S330, determine the n second attribute values corresponding to the target video on the n recommended attributes, and determine the n second attribute values corresponding to the target video on the n recommended attributes; Step S340, calculate the n first similarities between the target video and any case on the n recommended attributes according to the n first attribute values and the n second attribute values; Step S350, calculate the third similarity between the target video and any case according to the n first similarities and the weights of the n recommended attributes; Step S360, re-determine the target case from the M cases according to the third similarity and push it.

[0067] In Figure 2 case-based reasoning generally includes: retrieval, revision, and storage. The above steps adopt the case revision theory of case-based reasoning.

[0068] First, judge the user's degree of interest in the target case. For example: Step S410, confirm the viewing duration of the target case.

[0069] Step S420, calculate the ratio between the viewing duration and the total duration of the target case as the degree of interest.

[0070] For example, if the target case is 2 minutes, but in fact the user only watched for 10 seconds and then ended the viewing, then the calculated ratio is: 10 / 120.

[0071] The implementation method of this technical feature can include various variations. For example, the viewing duration data of the user can be obtained through video playback records, or the viewing duration can be automatically recorded through the built-in function of the video playback software. In addition, data such as the user's click behavior and pause times can be combined to further refine the calculation method of the degree of interest. Compared with the prior art, it pays more attention to the user's actual viewing behavior, can more accurately reflect the user's interest, and thus improves the accuracy of recommendations and user satisfaction.

[0072] In step S320, if the degree of interest is less than the threshold, for example, 12 / 120 < 1 / 2, it proves that the user is dissatisfied with the target case, so a new recommendation is made.

[0073] In steps S320 to S350, consider extracting the top n recommended attributes from N recommended attributes (weights), and use the n recommended attributes as the key recommended attributes to narrow the calculation space of the recommended attributes, reduce the features that cannot reflect the problem, and try to retain as many core features as possible. The specific operation is: recalculate the global similarity of the target video using the top n recommended attributes.

[0074] Finally, re-determine the target case from M cases based on the calculated third similarity and make recommendations.

[0075] This method can dynamically adjust the key recommended attributes based on weights, correct the cases, and improve the accuracy of food short video recommendations and user satisfaction.

[0076] Furthermore, this application also proposes to determine the target case from M cases according to the second similarity, including: Based on the second similarity, sort M cases in descending order to obtain a descending queue; Select the first case in the descending queue as the target case.

[0077] Through the above solution, this application can sort M cases in descending order to obtain a descending queue. The first case in the descending queue is the case most similar to the target video. Specifically, first compare all cases according to the second similarity between the target video and any case, and sort them from high to low according to the similarity. Then, select the first case from the sorted queue as the recommended target case. This method ensures that the recommended case has the highest similarity with the video currently browsed by the user in terms of recommended attributes, thus improving the accuracy of the recommendation.

[0078] The implementation method of the above technical features may include: first obtain the similarity value of each case by calculating the second similarity, and then sort these values to generate a descending queue. Sorting methods can use efficient sorting algorithms such as quicksort and mergesort. Selecting the first case in the queue as the target case can be achieved through simple array indexing operations.

[0079] Compared with the prior art, this application overcomes the problem of low recommendation accuracy in the prior art by sorting cases in descending order and selecting the most similar case as the recommended target. In this way, it can be ensured that the recommended food short video has the highest similarity with the video currently browsed by the user in multiple recommended attributes, thus improving user satisfaction and the effect of the recommendation system.

[0080] Furthermore, before step S110, the following steps are also included: Step S510, obtain multi-modal food data of the target area; Step S520: Construct multiple food short videos based on multi-modal food data; Step S530: Build a case library based on multiple food short videos.

[0081] By obtaining the multi-modal food data of the target area, it can include various forms of data such as images, videos, audio, and text. These data can better reflect the food characteristics of the target area. Based on these multi-modal data, multiple food short videos are constructed, and these short videos can more vividly and comprehensively display various attributes of the food. Then, based on these food short videos, a case library is built, which can ensure that the content in the case library is rich and representative.

[0082] For example, multi-modal food data can be obtained by crawling food pictures, videos, and relevant text descriptions about the target area on the Internet, or by actually shooting and recording the food production process in the target area. Then, using video editing tools, these data are integrated into multiple food short videos. Finally, these short videos are stored in the case library according to certain classification criteria for subsequent recommendation algorithms to use.

[0083] As Figure 2 shown, for the convenience of understanding, taking the recommendation of food short videos in Wuling Mountain Area as an example, the main purpose is to recommend the tourism resources in Wuling Mountain Area. For this purpose, a method for recommending regional food short videos based on case-based reasoning is provided. This method includes the following steps: (1) Extract food short videos about Wuling Mountain Area based on multi-modal technology; (2) Determine the first attribute values of food short videos in different recommended attributes based on preset recommended attributes; (3) Build a case library based on food short videos and their corresponding first attribute values; (4) Obtain the target video currently browsed by the user on the mobile phone; (5) Determine the second attribute value corresponding to the target video based on preset recommended attributes; (6) Calculate the local similarity between the target video and different food short videos on different recommended attributes based on the first attribute values of different recommended attributes and the second attribute value; (7) Calculate the weights of different recommended attributes using information entropy technology based on the first attribute values of food short videos in the case library on different recommended attributes; (8) Calculate the global similarity between the target video and different food short videos based on the local similarity and weights; (9) Search for the target case based on the global similarity; (10) Judge the degree of interest of the user in the target case; (11) Modify the target case based on the degree of interest; that is, select several recommended attributes with the highest weight values and recalculate the global similarity once according to steps (2) to (8). (12) Search for a target case again based on the global similarity.

[0084] (13) Take the current video as a case in the case library.

[0085] The method has at least the following beneficial effects: (1) Compared with the prior art that mainly relies on keywords and user tags to find target cases and ignores the recommended attributes of the video itself, resulting in low recommendation accuracy; based on the theory of case-based reasoning technology, this method first calculates the similarity between the video and the case in terms of recommended attributes, and then combines the weights of the recommended attributes to finally calculate the target case, which can better reflect the user's interest and the recommended value of the case itself, and improve the recommendation accuracy.

[0086] (2) Compared with the technical solution of assigning corresponding tags to each case, this method designs multiple recommended attributes, covering more recommendation directions. Each case has corresponding recommended attribute values on multiple recommended attributes, which also reflects the recommended value of the case on different recommended attributes, not limited to corresponding to a certain type of tag. In this way, the accuracy of food short video recommendation can be improved.

[0087] (3) Based on the theory of case-based reasoning, this application can dynamically adjust key recommended attributes based on weights to modify the case, so as to improve the accuracy of food short video recommendation and user satisfaction.

[0088] (4) Based on the attribute values of each case on different recommended attributes, this application adopts an information entropy calculation scheme to calculate the weights of different recommended attributes, strengthen the correlation between the calculated weights and the attribute values on different recommended attributes, improve the accuracy of the calculated weights, and finally improve the accuracy of food short video recommendation.

[0089] As Figure 4 shown, a device for recommending regional food short videos based on case-based reasoning is provided. The device includes: The data acquisition module 1100 is used to acquire the target video and the case library. Among them, there are M cases in the case library, and N first attribute values corresponding to each case on N recommended attributes; the case is a food short video of the target region; the target video is the currently browsed short video; N and M are positive integers greater than 1. The attribute value determination module 1200 is used to determine N second attribute values corresponding to the target video on N recommended attributes. The first similarity calculation module 1300 is configured to calculate N first similarities between the target video and any case on N recommended attributes according to N first attribute values and N second attribute values; The second similarity calculation module 1400 is configured to calculate the second similarity between the target video and any case according to N first similarities and the weights of N recommended attributes; The case selection module 1500 is configured to determine a target case from M cases according to the second similarity; The case push module 1600 is configured to push the food short video corresponding to the target case.

[0090] It should be noted that the case-based reasoning-based regional food short video recommendation device provided in this embodiment and the above-mentioned case-based reasoning-based regional food short video recommendation method are based on the same inventive concept. Therefore, the relevant content of the above-mentioned case-based reasoning-based regional food short video recommendation method also applies to the content of the case-based reasoning-based regional food short video recommendation device. Therefore, it will not be elaborated here.

[0091] As Figure 5 shown, an embodiment of the present application further provides an electronic device, which includes: At least one memory; At least one processor; At least one program; The program is stored in the memory, and the processor executes at least one program to implement the above-mentioned case-based reasoning-based regional food short video recommendation method of the present disclosure.

[0092] The electronic device may be any intelligent terminal including a mobile phone, a tablet computer, a personal digital assistant (PDA), an in-vehicle computer, etc.

[0093] The electronic device of the embodiment of the present application will be introduced in detail below.

[0094] The processor 1600 can be implemented by using a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement the technical solutions provided by the embodiments of the present invention; The memory 1700 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 1700 can store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 1700 and are called by the processor 1600 to execute the case-based reasoning method for recommending short videos of local cuisine in the embodiments of the present invention.

[0095] The input / output interface 1800 is used to implement information input and output. The communication interface 1900 is used to implement communication and interaction between this device and other devices. It can achieve communication through wired means (such as USB, network cable, etc.) or through wireless means (such as mobile network, WIFI, Bluetooth, etc.). The bus 2000 transmits information between various components of the device (such as the processor 1600, the memory 1700, the input / output interface 1800, and the communication interface 1900). Among them, the processor 1600, the memory 1700, the input / output interface 1800, and the communication interface 1900 are communicatively connected to each other inside the device through the bus 2000.

[0096] The embodiments of the present invention also provide a storage medium, which is a computer-readable storage medium. The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to cause a computer to execute the above-mentioned case-based reasoning method for recommending short videos of local cuisine.

[0097] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory can include high-speed random access memory, and can also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices.

[0098] In some embodiments, the memory optionally includes a memory remotely set relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0099] The embodiments described in the present invention are for more clearly illustrating the technical solutions of the embodiments of the present invention, and do not constitute a limitation to the technical solutions provided by the embodiments of the present invention. As can be known to those skilled in the art, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present invention are equally applicable to similar technical problems.

[0100] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation to the embodiments of the present invention, and may include more or fewer steps than those shown in the figures, or combine certain steps, or different steps.

[0101] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0102] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, and their appropriate combinations.

[0103] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of this application and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances, so that the embodiments of this application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0104] It should be understood that in this application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that there can be three relationships. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist at the same time. Among them, A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one (item) of the following" or its similar expression refers to any combination of these items, including any combination of single item (item) or plural items (items). For example, at least one (item) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0105] In several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of devices or units can be in electrical, mechanical or other forms.

[0106] 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 can be located in one place, or can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0107] In addition, in each embodiment of this application, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0108] When an integrated unit 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 technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions for causing an electronic device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of various embodiments of this application. The aforementioned storage medium includes: various media that can store programs, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0109] The above has specifically described the preferred implementation of the embodiments of this application. However, the embodiments of this application are not limited to the above-mentioned implementation manners. Those skilled in the art can also make various equivalent deformations or substitutions without departing from the spirit of the embodiments of this application. These equivalent deformations or substitutions are all included within the scope defined by the claims of the embodiments of this application.

Claims

1. A method for recommending short videos of local cuisine based on case-based reasoning, characterized in that, The method includes: Obtain a target video and a case library, where there are M cases in the case library, and N first attribute values corresponding to each case on N recommended attributes; the cases are short food videos of the target area; the target video is the currently viewed short video; N and M are positive integers greater than 1; Determine N second attribute values corresponding to the target video on the N recommended attributes; According to the N first attribute values and the N second attribute values, calculate N first similarities between the target video and any case on the N recommended attributes; According to the N first similarities and the weights of the N recommended attributes, calculate a second similarity between the target video and any case; Determine a target case from the M cases according to the second similarity; Push the short food video corresponding to the target case.

2. The case-based reasoning-based method for recommending short videos of local cuisine according to claim 1, characterized in that Before calculating the second similarity between the target video and any case according to the N first similarities and the weights of the N recommended attributes, it further includes: Calculate the weight of each recommended attribute: Among them, is the weight of the th recommended attribute, is the normalized value of the th case on the th recommended attribute, is the minimum value of the th case on the th recommended attribute, is the maximum value of the th case on the th recommended attribute, is the first attribute value of the th case on the th recommended attribute; is the information entropy of the th recommended attribute, is the natural logarithm function; The process of calculating the second similarity between the target video and any case includes: Among them, is the target video, is the target video and the th case at the th recommended attribute, the first similarity, is the target video and the th case of the second similarity.

3. The method for recommending short videos of local cuisine based on case-based reasoning according to claim 2, wherein After pushing the short food video corresponding to the target case, it further includes: Judge the degree of interest in the target case; If the degree of interest is less than the threshold, select n recommended attributes from the N recommended attributes in descending order of weight; n is less than N and n is a positive integer; Determine n second attribute values corresponding to the target video on the n recommended attributes, and determine n second attribute values corresponding to the target video on the n recommended attributes; According to the n first attribute values and the n second attribute values, calculate n first similarities between the target video and any case on the n recommended attributes; According to the n first similarities and the weights of the n recommended attributes, calculate a third similarity between the target video and any case; Redetermine the target case from the M cases according to the third similarity and push it.

4. The method for recommending short videos of local cuisine based on case-based reasoning according to claim 2, characterized in that, The determining the target case from the M cases according to the second similarity includes: Based on the second similarity, sort the M cases in descending order to obtain a descending queue; Select the first case in the descending queue as the target case.

5. The method for recommending short videos of local cuisine based on case-based reasoning according to claim 4, wherein The determining method of the degree of interest includes the following steps: Confirm the viewing duration of the target case; Calculate the ratio between the viewing duration and the total duration of the target case as the degree of interest.

6. The method for recommending short videos of local cuisine based on case-based reasoning according to claim 1, wherein The recommended attributes include: core content attribute, visual presentation attribute, narrative structure attribute, functional value attribute, communication value attribute; The core content attribute includes: local cuisine, street snacks, creative cuisine; The visual presentation attribute includes: camera language, color style; The narrative structure attribute includes: golden 5-second rule, information density; The functional value attribute includes: tutorial type, store visit type, cultural type; The communication value attribute includes: view count, like count.

7. The case-based reasoning method for short video recommendation of local cuisine according to claim 1, characterized in that, Before obtaining the target video and the case library, the method further includes: Obtain multimodal food data of the target area; Construct multiple food short videos based on the multimodal food data; Build the case library based on the multiple food short videos.

8. A short video recommendation device for local cuisine based on case-based reasoning, characterized in that, The device includes: A data acquisition module, configured to acquire a target video and a case library, where there are M cases in the case library, and N first attribute values corresponding to each case on N recommended attributes; the case is a food short video of the target area; the target video is the currently viewed short video; N and M are positive integers greater than 1; An attribute value determination module, configured to determine N second attribute values corresponding to the target video on the N recommended attributes; A first similarity calculation module, configured to calculate N first similarities between the target video and any case on the N recommended attributes according to the N first attribute values and the N second attribute values; A second similarity calculation module, configured to calculate a second similarity between the target video and any case according to the N first similarities and the weights of the N recommended attributes; A case selection module, configured to determine a target case from the M cases according to the second similarity; A case push module, configured to push the food short video corresponding to the target case.

9. An electronic device, characterized in that, It includes: At least one controller and a memory for communicatively connecting with the at least one controller; The memory stores instructions executable by the at least one controller, and the instructions are executed by the controller to enable the controller to execute the case-based reasoning method for recommending regional food short videos according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions for causing a computer to execute the case-based reasoning method for recommending regional food short videos according to any one of claims 1 to 7.

Images