阅读说明：本技术 一种基于人工智能技术对糖尿病人进行饮食结构推荐的方法 (Method for recommending dietary structure of diabetic patient based on artificial intelligence technology ) 是由 李锐 张晖 姜凯 于 2021-08-20 设计创作，主要内容包括：本发明提供了一种基于人工智能技术对糖尿病人进行饮食结构推荐的方法。这项技术需要将大量医院以及网络上可靠的数据进行收集处理搭建数据库,并将数据库存入云端。在对用户进行饮食推荐时,需要依赖物联网技术从各个终端获取用户的数据。在获取数据后,依靠人工智能模型对获取的数据进行整理和转化。根据整理和转化的数据,从数据库中找出类似的病例,用这些病例搭建机器学习和深度学习模型。根据模型和最优化理论得出最优的饮食推荐,并通过终端反馈给用户。(The invention provides a method for recommending a diet structure of a diabetic patient based on an artificial intelligence technology. The technology needs to collect and process reliable data of a large number of hospitals and networks to build a database, and the database is stored in the cloud. When a user is recommended to eat, the data of the user needs to be acquired from each terminal by means of the technology of the internet of things. After the data are acquired, the acquired data are sorted and converted by means of an artificial intelligence model. And finding similar cases from the database according to the sorted and converted data, and constructing machine learning and deep learning models by using the cases. And obtaining the optimal diet recommendation according to the model and the optimization theory, and feeding back to the user through a terminal.)

1. A method for recommending a dietary structure of a diabetic patient based on an artificial intelligence technology is characterized by comprising the following steps:

step 1: collecting a large amount of diets of diabetics and data of various physical characteristics from hospitals and credible websites to build a database;

step 2: when the user registers on the application, establishing an independent file for the user, wherein the file comprises various body data;

and step 3: collecting real-time body data of the user from the terminal equipment by using the technology of the Internet of things, and perfecting the file established in the step 2 according to the body data input by the user from the mobile phone and the data input by each equipment of the Internet of things;

and 4, step 4: converting the real-time body data collected in the step 2 by using machine learning and deep learning contents, building a logistic regression, LDA, QDA, Naive Bayes, SVM, Random Forest and Neural Network classification models by using data in a database when establishing whether a user has an abnormal state, and ensuring that each model has an optimal prediction effect by using a cross validation method when searching parameters for the models;

and 5: from the database, a case similar to the user's body data is found according to step 4.

Step 6: and (5) building a machine learning model, a deep learning model and an optimization model by using the data found in the step 5, and summarizing the results of a plurality of models into a dietary structure recommended to the user.

2. The method for dietary structure recommendation for diabetic patients based on artificial intelligence technology according to claim 1, wherein the database construction process is as follows:

each hospital mainly records the blood sugar values of the patient before and after meals every day, a blood sugar change index is constructed to represent the blood sugar change condition of the patient in the hospitalization period, for each time period, the last blood sugar value is used for subtracting the first blood sugar value and dividing the first blood sugar value by the number of days of difference of two blood sugar value measurement time, 6 blood sugar difference values can be obtained, then the 6 blood sugar difference values are weighted and averaged, the weight is the number of days of difference corresponding to each blood sugar difference value, and a is set₁,...,a₆For the calculated blood sugar difference values, six time intervals before, during and after meals, d₁,...,d₆The number of days for difference in each time period is determined as

Smaller C indicates more obvious blood sugar drop, and a change index C corresponds to a group of body data and recipes in the database.

3. The method for recommending a dietary structure of a diabetic patient based on artificial intelligence technology as claimed in claim 1, wherein said terminal device in step 3 includes but is not limited to a mobile phone, and further includes a weight scale and a treadmill connected to the internet of things.

4. The method for recommending dietary patterns of diabetic patients based on artificial intelligence technology according to claim 1, wherein the specific process of inducing the recommended dietary patterns of users in the step 6 is as follows: the blood sugar variation is taken as a dependent variable, the intake of various foods is taken as an independent variable, a decision tree model is built, leaf nodes with the most obvious blood sugar reduction are found out in the decision tree model, and then the optimal diet condition is summarized from the diet condition of the leaf node cases.

5. The method for recommending dietary patterns of diabetic patients based on artificial intelligence technology according to claim 1, wherein the specific process of inducing the recommended dietary patterns of users in the step 6 is as follows: building an objective function C-Q (V, W), wherein C is a blood sugar change index, V is various food intakes, W is various body data of a body, building Q can be various deep learning and machine learning models, then searching V corresponding to the minimum value of C by using an optimization method such as gradient descent and the like, and the V is various food intakes for the maximum blood sugar descent.

Technical Field

The invention relates to a method for recommending a diet structure of a diabetic patient based on an artificial intelligence technology, and belongs to the technical field of artificial intelligence and the Internet of things.

Background

Under the background of big data, artificial intelligence has wide application prospect, and machine learning and deep learning have great application potential as main technical means under the artificial intelligence framework. Machine learning and deep learning rely on the powerful computing power of a computer and learn the existing mass data by means of a model, so that the model has powerful decision-making and analysis capabilities. And the technology of the internet of things provides a more convenient way for acquiring and transmitting data. Currently, the combination of many fields with artificial intelligence and internet of things technology is still limited. A method for collecting data by using the Internet of things, building an artificial intelligence model by using the data and recommending diet of a diabetic patient by using an intelligent technology is lacked.

Disclosure of Invention

The invention aims to provide a method for recommending a dietary structure of a diabetic patient based on an artificial intelligence technology, which obtains optimal dietary recommendation according to a model and an optimization theory and feeds the optimal dietary recommendation back to the user through a terminal.

In order to achieve the purpose, the invention is realized by the following technical scheme:

a method for recommending a diet structure of a diabetic patient based on an artificial intelligence technology comprises the following steps:

step 1: collecting a large amount of diets of diabetics and data of various physical characteristics from hospitals and credible websites to build a database;

step 2: when the user registers on the application, establishing an independent file for the user, wherein the file comprises various body data;

and step 3: collecting real-time body data of the user from the terminal equipment by using the technology of the Internet of things, and perfecting the file established in the step 2 according to the body data input by the user from the mobile phone and the data input by each equipment of the Internet of things;

and 4, step 4: converting the real-time body data collected in the step 2 by using machine learning and deep learning contents, building a logistic regression, LDA, QDA, Naive Bayes, SVM, Random Forest and Neural Network classification models by using data in a database when establishing whether a user has an abnormal state, and ensuring that each model has an optimal prediction effect by using a cross validation method when searching parameters for the models;

and 5: from the database, a case similar to the user's body data is found according to step 4.

Step 6: and (5) building a machine learning model, a deep learning model and an optimization model by using the data found in the step 5, and summarizing the results of a plurality of models into a dietary structure recommended to the user.

Preferably, the database construction process is as follows:

each hospital mainly records the blood sugar values of the patient before and after meals every day, a blood sugar change index is constructed to represent the blood sugar change condition of the patient in the hospitalization period, for each time period, the last blood sugar value is used for subtracting the first blood sugar value and dividing the first blood sugar value by the number of days of difference of two blood sugar value measurement time, 6 blood sugar difference values can be obtained, then the 6 blood sugar difference values are weighted and averaged, the weight is the number of days of difference corresponding to each blood sugar difference value, and a is set₁,...,a⁶For the calculated blood sugar difference values, six time intervals before, during and after meals, d₁,...,d₆The number of days for difference in each time period is determined as

Smaller C indicates more obvious blood sugar drop, and a change index C corresponds to a group of body data and recipes in the database.

Preferably, the terminal device in step 3 includes, but is not limited to, a mobile phone, and further includes a weight scale and a treadmill connected to the internet of things.

Preferably, the specific process of inducing the user recommended dietary structure in the step 6 is as follows: the blood sugar variation is taken as a dependent variable, the intake of various foods is taken as an independent variable, a decision tree model is built, leaf nodes with the most obvious blood sugar reduction are found out in the decision tree model, and then the optimal diet condition is summarized from the diet condition of the leaf node cases.

Preferably, the specific process of inducing the user recommended dietary structure in the step 6 is as follows: building an objective function C-Q (V, W), wherein C is a blood sugar change index, V is various food intakes, W is various body data of a body, building Q can be various deep learning and machine learning models, then searching V corresponding to the minimum value of C by using an optimization method such as gradient descent and the like, and the V is various food intakes for the maximum blood sugar descent.

The invention has the advantages that: the method comprises the steps of receiving various body data of a user from each terminal in the Internet of things, matching the collected data with the body data of the user, finding out data similar to the body data of the user, establishing an artificial intelligence optimization model by using the found data, and inducing a recommended recipe for the user according to a model result.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention.

Fig. 1 is a schematic front view of an embodiment 1 of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

An application of diet structure recommendation for diabetes patients based on artificial intelligence technology. The specific content and steps are as follows:

step 1: a database is built by collecting a large number of diabetic diets and data of various physical characteristics from hospitals and trusted websites. From the currently available data, hospitals mainly record blood glucose values of patients before and after meals (six times in total), but blood glucose values of patients are not recorded every day, and we are more concerned about the variation of blood glucose of patients during hospitalization, so that a blood glucose variation index is constructed to represent the blood glucose variation of patients during hospitalization. For each time interval, subtracting the first blood glucose value from the last blood glucose value, dividing the difference between the measurement time of the two blood glucose values by the number of days to obtain 6 blood glucose difference values, and then carrying out weighted average on the 6 blood glucose difference values, wherein the weight is the number of days which is different from the number of days corresponding to each blood glucose difference value.

Step 2: when the user registers on the application, establishing an independent file for the user, wherein the file comprises various body data;

and step 3: collecting real-time body data of the user from the terminal equipment by using the technology of the Internet of things, and perfecting the file established in the step 2 according to the body data input by the user from the mobile phone and the data input by each equipment of the Internet of things;

and 4, step 4: converting the real-time body data collected in the step 2 by using machine learning and deep learning contents, building a logistic regression, LDA, QDA, Naive Bayes, SVM, Random Forest and Neural Network classification models by using data in a database when establishing whether a user has an abnormal state, and ensuring that each model has an optimal prediction effect by using a cross validation method when searching parameters for the models;

and 5: from the database, a case similar to the user's body data is found according to step 4.

Step 6: and (5) building a machine learning model, a deep learning model and an optimization model by using the data found in the step 5, and summarizing the results of a plurality of models into a dietary structure recommended to the user.