Cooking robot menu management method and system based on big data and medium

Abstract

The invention discloses a cooking robot menu management method and system based on big data and a medium, and relates to the technical field of cooking robot menu management. The cooking robot menu management method comprises the following steps: a data acquisition step: collecting various menu data through a web crawler technology; a data preprocessing step: cleaning the collected menu data; a data analysis and classification step: analyzing the preprocessed menu data by using a data mining algorithm; a menu recommendation and customization step: recommending a proper menu to the user by using a collaborative filtering algorithm or a content-based recommendation algorithm according to historical ordering records, health requirements and personal taste preference information of the user; and an intelligent cooking control step: the cooking robot receives the customized cooking instruction. According to the invention, the modules cooperatively and efficiently operate, links of data acquisition, processing, analysis, recommendation, cooking control and the like are closely linked, the system response is rapid, and the user operation experience and cooking efficiency are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of cross-border e-commerce security systems, and in particular to a cooking robot recipe management method, system and medium based on big data. Background Art

[0002] With the rapid development of science and technology and the continuous improvement of people's living standards, the field of smart home has made significant progress. As an important part of it, cooking robots have gradually entered people's daily lives. However, there are still many challenges and problems in its development process.

[0003] Traditional cooking robot recipe management often relies on pre-built fixed recipes, which are limited in number and difficult to update. These built-in recipes are usually determined during the robot development stage, and it is difficult to meet the increasingly diverse taste needs of users and the ever-changing dietary trends. For example, when new ingredient combinations or cooking styles become popular, traditional cooking robots cannot follow up in time, resulting in limited user experience.

[0004] In terms of data acquisition, cooking robots have previously lacked effective means of collecting big data. It is difficult to build a rich and comprehensive recipe library by relying solely on manual entry or sharing a small amount of recipe data with a few partners. This not only narrows the range of recipes that the robot can choose from, but also makes it impossible to mine valuable information from massive amounts of data, such as the popular matching trends of ingredients, the differences in taste preferences of users in different regions, etc., making it difficult to achieve intelligent recipe recommendations and personalized cooking services.

[0005] Traditional methods of analyzing and managing recipes are mostly based on simple classification or manual experience judgment. Lacking precise data mining and analysis technology, it is impossible to deeply analyze key factors such as the complex relationship between ingredients in recipes and the impact of cooking steps on dish quality. This results in traditional cooking robots being unable to cope with special user needs, such as adjusting the nutritional content of recipes according to personal health conditions and intelligently generating new recipes based on food inventory.

[0006] In addition, in terms of cooking process control, although the cooking robot can perform basic cooking operations according to the preset program, it lacks the ability to accurately monitor and dynamically adjust the real-time cooking status. For example, it is unable to flexibly adjust key parameters such as heat and cooking time according to factors such as the actual heating conditions of the ingredients and changes in moisture content, which can easily lead to unstable cooking effects and uneven food quality.

[0007] At the same time, with the acceleration of globalization and regional exchange and integration of catering culture, local special recipes emerge in an endless stream. Traditional cooking robots are difficult to adapt to this rapidly changing food culture environment, and are unable to timely integrate and effectively utilize new recipe resources around the world.

[0008] In summary, the existing cooking robot recipe management technology can no longer meet the modern consumers' pursuit of food diversity, personalization, and intelligent cooking experience. Therefore, it is of great practical significance to develop a cooking robot recipe management method, system, and medium based on big data, which can effectively integrate rich recipe data resources, use advanced data processing and analysis technology, realize intelligent recipe management and personalized cooking services, and promote the further development and popularization of cooking robots in the field of smart homes. Summary of the invention

[0009] The present invention proposes a cooking robot recipe management method, system and medium based on big data to solve the problems mentioned in the above-mentioned prior art.

[0010] In order to achieve the above purpose, the present invention adopts the following technical solution: a cooking robot recipe management method based on big data, comprising: Data collection steps: We collected various recipe data from food websites and cooking forums through web crawler technology, including dish names, ingredient lists, cooking steps, taste preferences, nutritional ingredients, and user evaluation information. We also established connections with the dish databases of major catering institutions to obtain the special recipes and actual cooking data of professional chefs and build a large recipe data set. The web crawler sets intelligent screening rules to screen high-value recipe data based on the keyword library and the popularity evaluation model. The popularity evaluation model integrates page views, likes, and comments. The popularity evaluation formula is: , where H is the popularity value, V is the page views, L is the number of likes, and C is the number of comments. is the weight coefficient; Data preprocessing steps: clean the collected recipe data, remove duplicate, erroneous, and incomplete data, delete incorrectly formatted text, correct typos in ingredient names, and supplement missing cooking steps; then standardize the data, unify ingredient names, measurement units, and cooking terms, and make the data format consistent. The cleaning process uses a text error correction algorithm based on natural language processing, combined with semantic understanding and vocabulary matching, to identify and correct erroneous information. The text error correction accuracy evaluation formula is as follows: ,in is the accuracy, To correct the correct amount of text, is the total number of texts to be corrected; Data analysis and classification steps: Use data mining algorithms to analyze the preprocessed recipe data, mine the ingredient association rules, taste similarities, and cooking difficulty level characteristics between different dishes, find out the ingredient combinations used together through association analysis algorithms, and use clustering algorithms to classify recipes into spicy, sweet and sour, and light categories according to taste preferences. At the same time, combine machine learning algorithms to predict the popularity of recipes, use Apriori algorithms to mine ingredient association rules, and set the confidence calculation formula for ingredient association rules: , where C is the confidence level, is the probability of ingredients X and Y appearing at the same time, is the probability of ingredient X appearing; Recipe recommendation and customization steps: Based on the user's historical ordering records, health needs, and personal taste preference information, use collaborative filtering algorithms or content-based recommendation algorithms to recommend suitable recipes to users; and allow users to customize recommended recipes according to their own needs, and the system automatically generates customized cooking instructions; the collaborative filtering algorithm uses a hybrid model based on user similarity and item similarity, comprehensively considering user rating behavior and recipe characteristics, and sets the user similarity calculation formula: ,in is the similarity between users u and v, and are the ratings of dishes given by users u and v respectively, and is the average rating of users u and v, and I is the set of dishes that are jointly rated; Intelligent cooking control steps: The cooking robot receives customized cooking instructions and automatically controls the amount of ingredients, cooking time, heat, and frequency of stir-frying. During the cooking process, it monitors the cooking status in real time, such as temperature, humidity, and doneness of ingredients, and automatically adjusts the cooking parameters based on the monitoring results. The cooking robot uses an electronic scale combined with a spiral feeder to control the amount of ingredients, and sets a heat adjustment formula: ,in is the adjusted temperature value, T is the current temperature, is the target temperature, K is the temperature adjustment coefficient, is the previous temperature value; Data update and optimization steps: After each cooking is completed, user feedback and actual cooking data are collected, and these data are integrated and analyzed with the original recipe data. The recipe data model is updated and optimized using the incremental learning algorithm, so that the recipes can continuously adapt to the needs of different users and changes in actual cooking conditions. The incremental learning algorithm adopts a model update strategy based on gradient descent, adjusts the model weight according to the new data, and balances the learning efficiency and model stability; the model weight update formula is: ,in is the updated model weight, is the original model weight, is the learning rate, is the loss function, which is used to optimize the model weights according to the new data.

[0011] Furthermore, the data collection step also includes conducting a copyright review of the collected recipe data, establishing a copyright information database, recording the copyright ownership, authorization scope and other information of each recipe, marking or deleting recipe data without copyright or infringing copyright, and the copyright review uses image recognition and text comparison technology to identify copyright logos and text descriptions in recipe images and compare them with the copyright database.

[0012] Furthermore, in the data analysis and classification steps, a multimodal data fusion analysis method is adopted. In addition to text data, multimedia data such as dish pictures and cooking videos are also included in the analysis. Specifically, image recognition technology is used to analyze the type, color, and plating information of ingredients in dish pictures. Video analysis technology is used to extract cooking skills and action sequence features in cooking videos. These multimodal features are combined with text data features to construct a more comprehensive recipe feature model. After multimodal fusion analysis, the comprehensive evaluation accuracy of the model is improved by more than 15%; image recognition uses a convolutional neural network CNN model, and video analysis uses a spatiotemporal interest point detection algorithm to extract key frames and action sequences.

[0013] Furthermore, in the recipe recommendation and customization step, when the user inputs special ingredients or has dietary taboos, the system uses a constraint satisfaction algorithm to screen and adjust the recommended recipes. Specifically: if the user is allergic to seafood, the system will automatically exclude recipes containing seafood, and highlight the dishes without seafood and that meet the user's other needs in the recommended recipes. The screening accuracy of the constraint satisfaction algorithm reaches more than 95%. The constraint satisfaction algorithm uses a depth-first search strategy to build a constraint condition tree to quickly screen recipes that meet the conditions.

[0014] Furthermore, in the intelligent cooking control step, the cooking robot uses a combination of an infrared temperature sensor and a humidity sensor to monitor the cooking status in real time, and uses a fuzzy control algorithm to adjust the cooking parameters according to the monitoring data, including when the temperature is too high and the humidity is low, the fuzzy control algorithm will appropriately reduce the heat and add a small amount of water. The fuzzy control algorithm uses a triangular membership function to determine the fuzzy relationship between temperature, humidity and cooking parameters, and obtains accurate adjustment amounts through fuzzy reasoning and clarification processing.

[0015] Furthermore, in the data preprocessing step, semantic understanding technology is used to parse the cooking steps in the recipe, and the steps described in natural language are converted into structured instructions. The semantic understanding technology is based on the recurrent neural network (RNN) model, which learns a large amount of cooking step texts and understands the relationship between cooking actions and ingredients.

[0016] Furthermore, a system for implementing the cooking robot recipe management method based on big data includes: Data collection module: used to collect recipe data from multiple channels through web crawlers and connect with catering organization databases to obtain professional recipe data. It has the function of scheduling data collection tasks and can collect data according to the set time interval and data volume requirements to ensure the continuity and integrity of data collection. The data transmission rate of the data collection module is not less than 1Mbps. The web crawler adopts a distributed architecture and multi-node parallel collection. Data preprocessing module: cleans and standardizes the collected data, uses data quality monitoring algorithms to evaluate the cleaning and standardization effects, and automatically starts the reprocessing process when the data quality does not meet the standards. The data processing efficiency of the data preprocessing module is not less than 1,000 data per hour. The data quality monitoring algorithm is based on data feature statistics and rule verification, and monitors data quality indicators in real time; Data analysis and classification module: Use data mining and machine learning algorithms to analyze and classify preprocessed data, build multiple recipe feature models, including ingredient association models, taste classification models, and popularity prediction models, and provide model training and optimization functions. The model can be automatically updated according to new data. The data analysis and classification module uses GPU accelerated computing; Recipe recommendation and customization module: recommends recipes to users based on user information and data analysis results and supports customization. It has the function of learning user preferences and continuously adjusts the recommendation strategy based on user feedback. The recommendation response time of the recipe recommendation and customization module is less than 5 seconds. The recipe recommendation and customization module uses a cache mechanism to store popular recipes and user-frequently used recipes. Intelligent cooking control module: receives cooking instructions of customized recipes and controls the cooking process of the cooking robot, including the placement of ingredients, cooking time, heat, and stir-frying operations. It also monitors the cooking status in real time and adjusts parameters. The cooking parameter control accuracy of the intelligent cooking control module meets the set requirements. The intelligent cooking control module adopts an embedded real-time operating system. Data update and optimization module: collects user feedback and actual data after cooking, updates and optimizes the recipe data model, uses data version management function to trace changes in recipe data of different versions, and the update cycle of the data update and optimization module can be flexibly set according to user usage frequency and data changes. The data update and optimization module uses incremental storage technology to reduce data storage redundancy.

[0017] Furthermore, the system also includes a user interaction module that provides a visual interface. Users can operate through a touch screen or voice input. The interface operation response time is less than 2 seconds, and multi-language display and switching are supported. The user interaction module uses voice recognition technology based on a deep learning model.

[0018] Furthermore, the data analysis and classification module incorporates data on external factors such as season, region, and festival when constructing a popularity prediction model, and adopts a hybrid model that combines multiple linear regression with a neural network.

[0019] Furthermore, a computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the cooking robot recipe management method based on big data.

[0020] Compared with the prior art, the present invention has the following beneficial effects: As for recipe resources, through multi-channel data collection and integration, a huge and dynamically updated recipe database has been built, covering a large amount of food information, greatly enriching the user's range of choices. Whether it is local specialties or emerging creative dishes, they can be easily obtained to meet the needs of diverse tastes.

[0021] In terms of personalized experience, based on in-depth analysis of user historical data, such as ordering records, taste preferences, health status, etc., we use precise recommendation algorithms to customize exclusive recipes for users. The recommendation matching rate is as high as over 90%, and the cooking instruction generation accuracy rate exceeds 98%, realizing truly personalized cooking services and fully demonstrating the user's unique eating style and needs.

[0022] The cooking process has been optimized. Based on intelligent cooking control technology, the cooking robot monitors the cooking status in real time and accurately adjusts parameters such as the amount of ingredients added (accuracy of ±5 grams), cooking time (accuracy of ±10 seconds), heat (10 levels of precise adjustment) and stir-fry frequency (10 - 50 times / minute), ensuring stable and consistent cooking results. The cooking success rate exceeds 95%, effectively improving the quality and taste stability of dishes.

[0023] The data-driven optimization mechanism enables the system to update and improve the recipe data model using an incremental learning algorithm based on user feedback and actual data after each cooking session, improving the overall evaluation accuracy by more than 10%, constantly adapting to changes in different users and actual cooking conditions, and achieving self-evolution and continuous optimization of recipe management.

[0024] From the perspective of overall system performance, each module works together efficiently, and the links of data collection, processing, analysis, recommendation and cooking control are closely connected. The system responds quickly. For example, the response time for recipe recommendations is less than 5 seconds, which greatly improves the user's operating experience and cooking efficiency. It has effectively promoted cooking robots to new heights in the field of smart homes and created a new way of convenient, intelligent and personalized cooking life for users. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a schematic block diagram of a cooking robot recipe management method based on big data proposed by the present invention. DETAILED DESCRIPTION

[0026] 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.

[0027] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0028] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. In addition, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, and it can be the internal connection of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0029] Reference Figure 1: A cooking robot recipe management method based on big data, comprising: Data collection steps: Use web crawler technology to collect various recipe data from food websites, cooking forums and other channels, including dish names, ingredient lists, cooking steps, taste preferences, nutritional ingredients, user reviews and other information. The collection frequency is at least once a week, and the amount of data collected each time is no less than 1,000. At the same time, establish connections with the dish databases of major catering institutions to obtain professional chefs’ special recipes and actual cooking data, such as ingredient usage ratios, cooking time and heat control parameters, etc. The completeness of data collection reaches more than 90% to build a rich and comprehensive recipe big data set. The web crawler sets intelligent screening rules to screen high-value recipe data based on the keyword library and heat evaluation model. The keyword library covers common ingredients, popular cooking techniques, etc. The heat evaluation model integrates page views, likes, comments and other indicators to ensure that the accuracy of collected data reaches more than 92%. The heat evaluation formula is: , where H is the popularity value, V is the page views, L is the number of likes, and C is the number of comments. It is a weight coefficient determined based on big data analysis and can be dynamically adjusted to suit the data characteristics of different platforms.

[0030] Data preprocessing steps: Clean the collected recipe data to remove duplicate, erroneous, and incomplete data, such as deleting incorrectly formatted text, correcting typos in ingredient names, and supplementing missing cooking steps. The accuracy of the cleaned data reaches over 95%; then standardize the data to unify ingredient names, measurement units, cooking terms, etc., so that the data format is standardized and consistent. The consistency of the standardized data reaches over 98%, which is convenient for subsequent analysis and management. The cleaning process uses a text correction algorithm based on natural language processing, combined with semantic understanding and vocabulary matching, to identify and correct erroneous information, and the text correction accuracy exceeds 96%. Suppose the text correction accuracy evaluation formula is: ,in is the accuracy, To correct the correct amount of text, The total number of texts to be corrected.

[0031] Data analysis and classification steps: Use data mining algorithms to analyze the preprocessed recipe data, and mine the characteristics of ingredient association rules, taste similarities, cooking difficulty levels, etc. between different dishes. For example, use association analysis algorithms to find out the combinations of ingredients that are often used together, and use clustering algorithms to classify recipes into different categories such as spicy, sweet and sour, and light according to taste preferences. At the same time, combine machine learning algorithms to predict the popularity of recipes, with a prediction accuracy of more than 85%, providing a basis for recipe recommendation and optimization. Use the Apriori algorithm to mine ingredient association rules, with the support threshold set to 0.3 and the confidence threshold set to 0.6, to efficiently mine frequent item sets and improve the accuracy of association analysis. Suppose the confidence calculation formula for ingredient association rules is: , where C is the confidence level, is the probability of ingredients X and Y appearing at the same time, is the probability of ingredient X appearing, and the reliability of the association rule is evaluated by this formula.

[0032] Recipe recommendation and customization steps: Based on the user's historical ordering records, health needs, personal taste preferences and other information, collaborative filtering algorithms or content-based recommendation algorithms are used to recommend suitable recipes to users, with a matching degree of more than 90%; and users are allowed to customize recommended recipes according to their own needs, such as adjusting the amount of ingredients, replacing ingredients, modifying cooking steps, etc. The system automatically generates customized cooking instructions, and the accuracy of cooking instructions is more than 98%, meeting the user's personalized cooking needs. The collaborative filtering algorithm uses a hybrid model based on user similarity and item similarity, comprehensively considering user rating behavior and recipe characteristics to improve recommendation accuracy. Suppose the user similarity calculation formula is: ,in is the similarity between users u and v, and are the ratings of dishes given by users u and v respectively, and is the average rating of users u and v, and I is the set of commonly evaluated dishes. This formula is used to quantify the similarity between users and provide a basis for recommendation.

[0033] Intelligent cooking control steps: The cooking robot receives customized cooking instructions and automatically controls cooking parameters such as the amount of ingredients put in, cooking time, heat, and stir-fry frequency. The control accuracy of the amount of ingredients put in is ±5 grams, and the control accuracy of the cooking time is ±10 seconds. The heat can be accurately adjusted within 10 gears, and the stir-fry frequency can be accurately adjusted between 10-50 times per minute. During the cooking process, the cooking status is monitored in real time, such as temperature, humidity, and degree of doneness of ingredients. The cooking parameters are automatically adjusted according to the monitoring results to ensure the stability and consistency of the cooking effect, and the cooking success rate reaches more than 95%. The cooking robot uses a high-precision electronic scale combined with a spiral feeder to control the amount of ingredients put in. The accuracy of the electronic scale is 0.1 grams, and the speed error of the spiral feeder is less than 2%. Suppose the heat adjustment formula is: ,in is the adjusted temperature value, T is the current temperature, is the target temperature, K is the temperature adjustment coefficient, is the previous heat value. The heat size is adjusted dynamically according to the temperature change based on this formula.

[0034] Data update and optimization steps: After each cooking is completed, collect user feedback and evaluation, actual cooking data, such as actual cooking time, actual amount of ingredients used, user satisfaction with taste, etc., integrate and analyze these data with the original recipe data, and use the incremental learning algorithm to update and optimize the recipe data model, so that the recipe can continuously adapt to the needs of different users and changes in actual cooking conditions. The comprehensive evaluation accuracy of the updated recipe data model is improved by more than 10%. The incremental learning algorithm adopts a model update strategy based on gradient descent, adjusts the model weight according to the new data, and balances learning efficiency and model stability. Suppose the model weight update formula is: ,in is the updated model weight, is the original model weight, is the learning rate, is the loss function, which is used to optimize the model weights according to the new data.

[0035] The present invention also includes, in the data collection step, conducting a copyright review of the collected recipe data, establishing a copyright information database, recording the copyright ownership, authorization scope, and other information of each recipe, marking or deleting recipe data without copyright or infringement, and achieving a copyright review accuracy of more than 98%, ensuring the legality and compliance of the recipe data. The copyright review uses image recognition and text comparison technology to identify the copyright logo and text description in the recipe image, and compares it with the copyright database, with an identification accuracy of 99%.

[0036] In the data analysis and classification steps, the present invention adopts a multimodal data fusion analysis method. In addition to text data, multimedia data such as dish pictures and cooking videos are also included for analysis. For example, the type, color, and plating of ingredients in dish pictures are analyzed by image recognition technology, and the cooking skills, action sequence and other features in cooking videos are extracted by video analysis technology. These multimodal features are combined with text data features to build a more comprehensive recipe feature model. The comprehensive evaluation accuracy of the model after multimodal fusion analysis is improved by more than 15%. Image recognition uses a convolutional neural network (CNN) model, and the training set contains 100,000 dish pictures, with a recognition accuracy of 95%; video analysis uses a spatiotemporal interest point detection algorithm to extract key frames and action sequences, with an analysis accuracy of 90%.

[0037] In the recipe recommendation and customization step of the present invention, when the user inputs special ingredients or has dietary taboos, the system uses a constraint satisfaction algorithm to screen and adjust the recommended recipes. For example, if the user is allergic to seafood, the system will automatically exclude recipes containing seafood and highlight dishes without seafood that meet other user requirements in the recommended recipes. The screening accuracy of the constraint satisfaction algorithm reaches more than 95%, ensuring the user's dietary safety and health. The constraint satisfaction algorithm adopts a depth-first search strategy to construct a constraint condition tree, quickly screen recipes that meet the conditions, and improve search efficiency by 30%.

[0038] In the intelligent cooking control step of the present invention, the cooking robot uses a combination of an infrared temperature sensor and a humidity sensor to monitor the cooking status in real time, and uses a fuzzy control algorithm to adjust the cooking parameters according to the monitoring data. For example, when the temperature is too high and the humidity is low, the fuzzy control algorithm will appropriately reduce the heat and add a small amount of water. The fuzzy control algorithm has an accuracy rate of more than 90% in adjusting the cooking parameters, which improves the intelligent level of the cooking process. The fuzzy control algorithm uses a triangular membership function to determine the fuzzy relationship between temperature, humidity and cooking parameters, and obtains accurate adjustment amounts through fuzzy reasoning and clarification processing.

[0039] In the data preprocessing step, the present invention uses semantic understanding technology to parse the cooking steps in the recipe, converting the steps described in natural language into structured instructions, with a parsing accuracy of more than 95%, which facilitates the cooking robot to perform cooking operations more accurately. The semantic understanding technology is based on a recurrent neural network (RNN) model, which learns a large number of cooking step texts and understands the relationship between cooking actions and ingredients.

[0040] The present invention also discloses a cooking robot recipe management system based on big data, comprising: Data collection module: used to collect recipe data from multiple channels through web crawlers and connect with catering organization databases to obtain professional recipe data. It has data collection task scheduling function and can collect data according to the set time interval and data volume requirements to ensure the continuity and integrity of data collection. The data transmission rate of the data collection module is not less than 1Mbps. The web crawler adopts a distributed architecture and multi-node parallel collection, which improves the collection efficiency by more than 50%.

[0041] Data preprocessing module: cleans and standardizes the collected data, uses data quality monitoring algorithms to evaluate the cleaning and standardization effects, and automatically starts the reprocessing process when the data quality does not meet the standards. The data processing efficiency of the data preprocessing module is not less than 1,000 data per hour. The data quality monitoring algorithm is based on data feature statistics and rule verification, and monitors data quality indicators in real time, such as data integrity, accuracy, consistency, etc.

[0042] Data analysis and classification module: Use data mining and machine learning algorithms to analyze and classify preprocessed data, build multiple recipe feature models, such as ingredient association model, taste classification model, popularity prediction model, etc., and provide model training and optimization functions. The model training time does not exceed 24 hours, and the model can be automatically updated according to new data. The data analysis and classification module uses GPU accelerated computing to shorten the model training time by more than 40%.

[0043] Recipe recommendation and customization module: It recommends recipes to users based on user information and data analysis results and supports customization. It has the function of learning user preferences and can continuously adjust the recommendation strategy based on user feedback. The recommendation response time of the recipe recommendation and customization module is less than 5 seconds. The recipe recommendation and customization module uses a cache mechanism to store popular recipes and user-frequently used recipes, which improves the recommendation response speed by more than 60%.

[0044] Intelligent cooking control module: Receives cooking instructions from customized recipes and controls the cooking process of the cooking robot, including the placement of ingredients, cooking time, heat, stir-frying and other operations. It also monitors the cooking status in real time and adjusts parameters. The cooking parameter control accuracy of the intelligent cooking control module meets the set requirements. The intelligent cooking control module uses an embedded real-time operating system to ensure the timeliness and stability of cooking control, and the task response time is less than 10 milliseconds.

[0045] Data update and optimization module: collect user feedback and actual data after cooking, update and optimize the recipe data model, use data version management function to trace the changes of recipe data of different versions, and the update cycle of data update and optimization module can be flexibly set according to user usage frequency and data changes, and the shortest can be 1 day. The data update and optimization module uses incremental storage technology to reduce data storage redundancy and improve data update efficiency by more than 30%.

[0046] In the present invention, the system also includes a user interaction module, which provides a visual interface. Users can operate through touch screen, voice input, etc. The interface operation response time is less than 2 seconds, and it supports multi-language display and switching, which is convenient for different user groups. The user interaction module uses voice recognition technology based on a deep learning model, with a recognition accuracy of more than 98%, and supports multiple dialects and foreign language recognition.

[0047] In the present invention, the data analysis and classification module incorporates external factor data such as season, region, and festival when constructing a popularity prediction model, and adopts a hybrid model combining multivariate linear regression and neural network to improve the accuracy and adaptability of the prediction model. The model's accuracy in predicting the popularity of recipes in different regions and seasons is improved by more than 20%.

[0048] The present invention also discloses a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the cooking robot recipe management method based on big data are implemented.

[0049] The above are only preferred specific implementation modes of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solutions and inventive concepts of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A cooking robot recipe management method based on big data, characterized in that: include: Data collection steps: We collected various recipe data from food websites and cooking forums through web crawler technology, including dish names, ingredient lists, cooking steps, taste preferences, nutritional ingredients, and user evaluation information. We also established connections with the dish databases of major catering institutions to obtain the special recipes and actual cooking data of professional chefs and build a large recipe data set. The web crawler sets intelligent screening rules to screen high-value recipe data based on the keyword library and the popularity evaluation model. The popularity evaluation model integrates page views, likes, and comments. The popularity evaluation formula is: , where H is the popularity value, V is the page views, L is the number of likes, and C is the number of comments. is the weight coefficient; Data preprocessing steps: clean the collected recipe data, remove duplicate, erroneous, and incomplete data, delete incorrectly formatted text, correct typos in ingredient names, and supplement missing cooking steps; then standardize the data, unify ingredient names, measurement units, and cooking terms, and make the data format consistent. The cleaning process uses a text error correction algorithm based on natural language processing, combined with semantic understanding and vocabulary matching, to identify and correct erroneous information. The text error correction accuracy evaluation formula is as follows: ,in is the accuracy, To correct the correct amount of text, is the total number of texts to be corrected; Data analysis and classification steps: Use data mining algorithms to analyze the preprocessed recipe data, mine the ingredient association rules, taste similarities, and cooking difficulty level characteristics between different dishes, find out the ingredient combinations used together through association analysis algorithms, and use clustering algorithms to classify recipes into spicy, sweet and sour, and light categories according to taste preferences. At the same time, combine machine learning algorithms to predict the popularity of recipes, use Apriori algorithms to mine ingredient association rules, and set the confidence calculation formula for ingredient association rules: , where C is the confidence level, is the probability of ingredients X and Y appearing at the same time, is the probability of ingredient X appearing; Recipe recommendation and customization steps: Based on the user's historical ordering records, health needs, and personal taste preference information, use collaborative filtering algorithms or content-based recommendation algorithms to recommend suitable recipes to users; and allow users to customize recommended recipes according to their own needs, and the system automatically generates customized cooking instructions; the collaborative filtering algorithm uses a hybrid model based on user similarity and item similarity, comprehensively considering user rating behavior and recipe characteristics, and sets the user similarity calculation formula: ,in is the similarity between users u and v, and are the ratings of dishes given by users u and v respectively, and is the average rating of users u and v, and I is the set of dishes that are jointly rated; Intelligent cooking control steps: The cooking robot receives customized cooking instructions and automatically controls the amount of ingredients, cooking time, heat, and frequency of stir-frying. During the cooking process, it monitors the cooking status in real time, such as temperature, humidity, and doneness of ingredients, and automatically adjusts the cooking parameters based on the monitoring results. The cooking robot uses an electronic scale combined with a spiral feeder to control the amount of ingredients, and sets a heat adjustment formula: ,in is the adjusted temperature value, T is the current temperature, is the target temperature, K is the temperature adjustment coefficient, is the previous temperature value; Data update and optimization steps: After each cooking is completed, user feedback and actual cooking data are collected, and these data are integrated and analyzed with the original recipe data. The recipe data model is updated and optimized using an incremental learning algorithm, so that the recipes can continuously adapt to the needs of different users and changes in actual cooking conditions. The incremental learning algorithm adopts a model update strategy based on gradient descent, adjusts the model weight according to the new data, and balances learning efficiency and model stability; set up Model weight update formula: ,in is the updated model weight, is the original model weight, is the learning rate, is the loss function, which is used to optimize the model weights according to the new data.

2. The cooking robot recipe management method based on big data according to claim 1 is characterized in that: The data collection step also includes conducting a copyright review of the collected recipe data, establishing a copyright information database, recording the copyright ownership, authorization scope and other information of each recipe, marking or deleting recipe data without copyright or infringing copyright. The copyright review uses image recognition and text comparison technology to identify copyright logos and text descriptions in recipe images and compare them with the copyright database.

3. The cooking robot recipe management method based on big data according to claim 1 is characterized in that: In the data analysis and classification steps, a multimodal data fusion analysis method is adopted. In addition to text data, multimedia data such as dish pictures and cooking videos are also included for analysis. Specifically, image recognition technology is used to analyze the type, color, and plating information of ingredients in dish pictures. Video analysis technology is used to extract cooking skills and action sequence features in cooking videos. These multimodal features are combined with text data features to build a more comprehensive recipe feature model. After multimodal fusion analysis, the comprehensive evaluation accuracy of the model is improved by more than 15%; image recognition uses a convolutional neural network CNN model, and video analysis uses a spatiotemporal interest point detection algorithm to extract key frames and action sequences.

4. The cooking robot recipe management method based on big data according to claim 1 is characterized in that: In the recipe recommendation and customization step, when the user inputs special ingredients or has dietary taboos, the system uses a constraint satisfaction algorithm to screen and adjust the recommended recipes. Specifically: if the user is allergic to seafood, the system will automatically exclude recipes containing seafood and highlight dishes without seafood that meet the user's other needs in the recommended recipes. The screening accuracy of the constraint satisfaction algorithm reaches more than 95%. The constraint satisfaction algorithm uses a depth-first search strategy to build a constraint condition tree to quickly screen recipes that meet the conditions.

5. The cooking robot recipe management method based on big data according to claim 1 is characterized in that: In the intelligent cooking control step, the cooking robot uses a combination of an infrared temperature sensor and a humidity sensor to monitor the cooking status in real time, and uses a fuzzy control algorithm to adjust the cooking parameters according to the monitoring data, including when the temperature is too high and the humidity is low, the fuzzy control algorithm will appropriately reduce the heat and add a small amount of water. The fuzzy control algorithm uses a triangular membership function to determine the fuzzy relationship between temperature, humidity and cooking parameters, and obtains accurate adjustment amounts through fuzzy reasoning and clarification processing.

6. The cooking robot recipe management method based on big data according to claim 1 is characterized in that: In the data preprocessing step, semantic understanding technology is used to parse the cooking steps in the recipe and convert the steps described in natural language into structured instructions. The semantic understanding technology is based on the recurrent neural network (RNN) model, which learns a large amount of cooking step texts and understands the relationship between cooking actions and ingredients.

7. A system for implementing the cooking robot recipe management method based on big data as described in any one of claims 1 to 6, characterized in that: include: Data collection module: used to collect recipe data from multiple channels through web crawlers and connect with catering organization databases to obtain professional recipe data. It has the function of scheduling data collection tasks and can collect data according to the set time interval and data volume requirements to ensure the continuity and integrity of data collection. The data transmission rate of the data collection module is not less than 1Mbps. The web crawler adopts a distributed architecture and multi-node parallel collection. Data preprocessing module: cleans and standardizes the collected data, uses data quality monitoring algorithms to evaluate the cleaning and standardization effects, and automatically starts the reprocessing process when the data quality does not meet the standards. The data processing efficiency of the data preprocessing module is not less than 1,000 data per hour. The data quality monitoring algorithm is based on data feature statistics and rule verification, and monitors data quality indicators in real time; Data analysis and classification module: Use data mining and machine learning algorithms to analyze and classify preprocessed data, build multiple recipe feature models, including ingredient association models, taste classification models, and popularity prediction models, and provide model training and optimization functions. The model can be automatically updated according to new data. The data analysis and classification module uses GPU accelerated computing; Recipe recommendation and customization module: recommends recipes to users based on user information and data analysis results and supports customization. It has the function of learning user preferences and continuously adjusts the recommendation strategy based on user feedback. The recommendation response time of the recipe recommendation and customization module is less than 5 seconds. The recipe recommendation and customization module uses a cache mechanism to store popular recipes and user-frequently used recipes. Intelligent cooking control module: receiving cooking instructions of customized recipes, controlling the cooking process of the cooking robot, including the placing of ingredients, cooking time, heat, and stir-frying operations, and monitoring the cooking status in real time and adjusting parameters. The cooking parameter control accuracy of the intelligent cooking control module meets the requirements of claim 1, and the intelligent cooking control module adopts an embedded real-time operating system; Data update and optimization module: collects user feedback and actual data after cooking, updates and optimizes the recipe data model, uses data version management function to trace changes in recipe data of different versions, and the update cycle of the data update and optimization module can be flexibly set according to user usage frequency and data changes. The data update and optimization module uses incremental storage technology to reduce data storage redundancy.

8. The system of the cooking robot recipe management method based on big data according to claim 7 is characterized in that: The system also includes a user interaction module that provides a visual interface. Users can operate through a touch screen or voice input. The interface operation response time is less than 2 seconds, and it supports multi-language display and switching. The user interaction module uses voice recognition technology based on a deep learning model.

9. The system of the cooking robot recipe management method based on big data according to claim 7 is characterized in that: When constructing a popularity prediction model, the data analysis and classification module incorporates data on external factors such as season, region, and festival, and adopts a hybrid model combining multiple linear regression and neural network.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the cooking robot recipe management method based on big data described in any one of claims 1-5 are implemented.

Images