Machine learning-based feeding ability monitoring system for senile dementia patient

Through the machine learning-based dietary ability monitoring system for Alzheimer's patients, using video surveillance and sensors to collect data, build a food analysis model, which solves the problem that the existing system cannot fully analyze the food data, and realizes scientific adjustment of the food plan and the burden of nursing staff. The reduction of food safety and enthusiasm of patients.

CN119069077APending Publication Date: 2024-12-03THE FIRST AFFILIATED HOSPITAL OF ZHENGZHOU UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411167492.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-23
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The existing eating monitoring system for patients with Alzheimer's disease cannot fully obtain food data and cannot analyze and feedback the patients' eating ability, resulting in the inability to adjust the eating plan, which increases the workload of nursing staff.

Method used

Design a machine learning-based monitoring system for eating ability of patients with Alzheimer's disease, including data collection, transmission, storage, processing and analysis modules, collecting patient eating data through video monitoring and sensors, constructing a food analysis model, analyzing patients' eating situation in real time, and adjusting Meal plan to reduce the burden on caregivers.

Benefits of technology

It realizes comprehensive monitoring and analysis of the patient's eating process, improves the safety and enthusiasm of food, reduces the workload of nursing staff, and ensures the scientificity and safety of the food plan.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119069077A_ABST
    Figure CN119069077A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of monitoring, and discloses a machine learning-based senile dementia patient feeding ability monitoring system, which is characterized in that personal information, food data and feeding data of a patient are acquired and analyzed, and the analysis process comprises the steps of constructing a feeding analysis model to output feeding ability and food favorite values of the patient; the feeding ability and food preference of the patient can be known, the feeding plan of the patient can be effectively adjusted while the feeding process of the patient is monitored in real time, the safety and enthusiasm of the patient in the feeding process are improved, and the workload of nursing personnel is effectively reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of monitoring technology, and in particular to a system for monitoring the eating ability of patients with Alzheimer's disease based on machine learning. Background Art

[0002] Alzheimer's disease is a progressive neurodegenerative disease. The most common symptoms are memory loss and a decline in the patient's ability to perform daily activities. Therefore, caregivers are usually required to look after Alzheimer's patients 24 hours a day to prevent the patients from getting into danger.

[0003] When caring for patients with Alzheimer's disease, the patient's diet is an aspect that requires special attention. Due to the patient's memory loss and the decline in daily self-care ability, it is extremely risky for the patient to arrange his daily diet and eating by himself. At present, there are related monitoring products to monitor the patient's eating situation, but they are mainly through video monitoring, and the monitoring data is transmitted to the patient to achieve remote supervision, so as to timely identify dangerous situations that occur during the patient's eating process. However, the above products still have certain defects, mainly in that the data obtained from the patient's eating process is not comprehensive enough, and it is impossible to analyze and feedback the patient's eating ability to adjust the eating plan, and it cannot meet the requirements of reducing the patient's workload. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide an eating ability monitoring system for patients with Alzheimer's disease based on machine learning, so as to realize multi-faceted integration of patients' eating-related data. While safely monitoring the patients' eating process, the patients' eating ability can be obtained, and the patients' eating plans can be adjusted accordingly, thereby reducing the burden on caregivers.

[0005] The present invention discloses a system for monitoring the eating ability of patients with Alzheimer's disease based on machine learning, characterized in that the system comprises a data acquisition module, a data transmission module, a data storage module, a data processing and analysis module, and a display module;

[0006] The data collection module is used to collect patient personal information, patient food data, and patient eating data;

[0007] The data transmission module is used to transmit the data collected by the data collection module to the data storage module;

[0008] The data storage module is used to store the data collected by the data collection module;

[0009] The data processing and analysis module includes a data processing unit, a model training unit and a data analysis unit; the data processing unit is used to retrieve the data collected by the data collection module from the data storage module for data processing to obtain processed data; the data analysis unit is used to input the processed data into the eating analysis model and output the analysis results through the eating analysis model; the eating analysis model is obtained by training the model training unit based on the machine learning algorithm;

[0010] The data transmission module is also used to transmit the output information of the data processing and analysis module to the display module, and the user adjusts the patient's eating plan according to the analysis results of the model training unit.

[0011] Further, the data acquisition module includes an input acquisition unit, a video monitoring unit and a sensor unit;

[0012] The input collection unit is used to provide an input terminal for the user to input information, and collect the information input by the user, including collecting the patient's personal information and the food information of the patient's eating plan input by the user; wherein the food information in the eating plan input by the user is used as the first food data;

[0013] The video monitoring unit is used to collect food data of the patient's eating plan and the patient's eating data during the eating process; the food image data in the eating plan collected by the video monitoring unit is used as the second food data, and the patient's eating plan food data includes the first food data and the second food data; the food image of the actual eating collected by the video monitoring unit is used as the first eating data;

[0014] The sensor unit is used to collect the patient's actual eating state during the eating process through the sensor and use it as the second eating data. The patient's actual eating data includes the first eating data and the second eating data.

[0015] Furthermore, the data processing unit performs data processing on the data collected by the data collection module, including performing image recognition processing on the second food data and the first eating data respectively.

[0016] Furthermore, the patient's personal information includes the patient's age, gender, food taste, food taboos, chewing ability, and disease diagnosis history;

[0017] The first food data includes food names and dish names input by the user according to the patient's eating plan;

[0018] The second food data includes food name and dish name;

[0019] The first eating data includes the name of the food and dish eaten, eating time, eating speed, and eating degree;

[0020] The second eating data includes the tilt angle and weight of the tableware.

[0021] Furthermore, the system also includes a control module;

[0022] Before the patient eats, the data analysis unit determines the similarity value between the first food data and the second food data. When it is determined that the similarity value between the first food data and the second food data is greater than or equal to a preset threshold, the control module executes an operation to allow the patient to eat, and uses the second food data as input data for an eating analysis model.

[0023] Furthermore, the system further comprises a prompt module, and the prompt module is used to send prompt information to the user;

[0024] When it is determined that the similarity value between the first food data and the second food data is lower than a preset threshold, the prompt module sends a prompt message to the user, wherein the prompt message includes prompting the user to verify whether the food that the patient currently needs to eat meets the eating requirements.

[0025] Furthermore, the sensor unit includes a temperature detector, which detects the temperature of food that the patient is going to eat; the food temperature is obtained before the operation of allowing the patient to eat is performed, and when it is determined that the food temperature is lower than the preset food eating temperature, the operation of allowing the patient to eat is performed.

[0026] Furthermore, the patient's eating data is data collected in real time. The data analysis unit obtains the patient's eating data in real time, and performs real-time analysis based on the patient's eating data to determine whether the current eating data triggers an alarm event. When an alarm event is triggered, an alarm is issued to the user through the prompt module.

[0027] Furthermore, the model training unit input end receives processed data and outputs analysis results through the eating analysis model, specifically: the processed data is the patient's personal information and the first eating data, and the analysis results include an eating ability score and a preference value.

[0028] Furthermore, the machine learning algorithm includes a random forest algorithm, and each decision tree in the random forest model outputs a predicted value of the eating ability score and the preference value for each processed data, and the final output is a weighted average of the prediction results of all decision trees.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] The present invention collects and analyzes the patient's personal information, food data and eating data, wherein the analysis process includes constructing an eating analysis model to output the patient's eating ability and food preference value, so that the patient can understand the patient's eating ability and food preference, and can effectively adjust the patient's eating plan while achieving real-time monitoring of the patient's eating process, thereby improving the safety and enthusiasm of the patient's eating process and effectively reducing the workload of caregivers. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a structural schematic diagram of a system for monitoring the eating ability of patients with Alzheimer's disease based on machine learning disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0032] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments.

[0033] Embodiment 1

[0034] The present invention discloses a system for monitoring the eating ability of patients with Alzheimer's disease based on machine learning, and its structure is shown in the following figure: Figure 1 As shown, the system includes a data acquisition module, a data transmission module, a data storage module, a data processing and analysis module, and a display module;

[0035] The data collection module is used to collect patient personal information, patient food data, and patient eating data, and serves as data collected by the data collection module;

[0036] The data transmission module is used to transmit the data collected by the data acquisition module to the data storage module;

[0037] The data storage module is used to store the data collected by the data collection module;

[0038] The data processing and analysis module is used to process, analyze and monitor the patient's eating ability with respect to the data collected by the data acquisition module; the data processing and analysis module includes a data processing unit, a data analysis unit and a model training unit; the data processing unit is used to retrieve the data collected by the data acquisition module from the data storage module for data processing to obtain processed data; the data analysis unit is used to analyze and perform operations on the data collected by the data acquisition module, and the model training unit includes an eating analysis module, the eating analysis module input end receives the processed data, and the eating analysis model output end outputs the analysis result, wherein the eating analysis model is trained by the model training unit based on the machine learning algorithm.

[0039] The data transmission module is also used to transmit the analysis result to the display module, which displays the first analysis result, and the user adjusts the patient's eating plan according to the analysis result.

[0040] In the embodiment of the present invention, the data acquisition module processes the data collected by the data acquisition module, including data cleaning and data standardization processing operations.

[0041] Furthermore, the data acquisition module includes an input acquisition unit, a video monitoring unit and a sensor unit.

[0042] The input collection unit is used to provide an input end for the user to input information, and to collect the information input by the user, including collecting the patient's personal information input by the user and the food information for formulating a meal plan for the patient; wherein the food information in the meal plan input by the user is used as the first food data.

[0043] Furthermore, the patient's personal information includes but is not limited to the patient's age, gender, food taste, food taboos, chewing ability, and disease diagnosis history. The first food data includes but is not limited to the food names and dish names input by the user according to the patient's eating plan.

[0044] The video monitoring unit is used to collect food image data for formulating a meal plan for the patient, as well as the patient's meal data during the meal process; wherein the food image data in the meal plan collected by the video monitoring unit is used as the second food data, and the above-mentioned patient food data includes the first food data and the second food data.

[0045] Furthermore, the second food data includes food name and dish name;

[0046] The actual eating image data collected by the video monitoring unit is used as the first eating data.

[0047] The sensor unit is used to collect the patient's eating data during the eating process through the sensor and use it as the second eating data. The patient's eating data includes the first eating data and the second eating data.

[0048] Furthermore, the data processing unit performs data processing on the data collected by the data collection module, including performing image recognition processing on the second food data and the first eating data respectively.

[0049] Furthermore, the first eating data includes the name of the food and dish, eating time, eating speed, and eating degree; the second eating data includes the tilt angle of the tableware, weight,

[0050] Specifically, the user in the embodiment of the present invention refers to a person who is related to taking care of the patient, including but not limited to caregivers and family members of the patient. The patient's eating plan refers to the eating plan arranged for the patient according to the doctor's advice and the patient's eating ability. The eating plan can be in units of days, weeks, etc. Correspondingly, when the user inputs relevant information according to the patient's eating plan, the eating plan information can be input in units of days or weeks. The food name includes but is not limited to the fruits, beverages, ingredients of the dishes, and the names of the dishes that the patient wants to eat. Furthermore, the dish name includes the taste of the dish, such as light and non-spicy steamed fish, slightly spicy braised fish, etc. The taste of the dish can also include the content of seasonings such as salt, sugar, etc. The present invention does not specifically limit the taste representation contained in the dish name.

[0051] The processing unit performs image recognition processing on the second food data to obtain the second food data including food names and dish names. Specifically, the video monitoring unit photographs the food, dishes and food to be eaten by the patient in the patient's meal plan placed in the area around the patient, and the data processing unit recognizes the photographed food pictures based on image recognition technology to identify the specific food names and dish names.

[0052] The processing unit performs image recognition processing on the first eating data, and obtains the first eating data including but not limited to the name of the food and dish that the patient wants to eat, the eating time, the eating speed, the eating degree, and the tilt angle of the tableware, and the second eating data includes the tilt angle and weight of the tableware.

[0053] Specifically, in an embodiment of the present invention, the second food data and the first eating data are obtained by image recognition based on the image data of the patient's eating process monitored by the video monitoring unit through image recognition technology. The name of the food and the name of the dish are recognized in the same process as the first food data. The eating time is obtained based on the time corresponding to the start eating image and the end eating image in the image data; the eating speed is obtained based on the change in the height of the food; the eating degree is calculated based on the height and weight of the food at the start and end of the meal, wherein the change in food weight needs to be obtained in combination with the second eating data collected by the sensor, and the eating degree specifically refers to the patient's food intake; the tableware inclination angle is obtained based on the recognition of the shape of the tableware.

[0054] Furthermore, the system also includes a control module.

[0055] Before the patient eats, the data analysis unit determines the similarity between the first food data and the second food data. When the similarity is greater than or equal to a preset threshold, the control module allows the patient to eat and uses the second food data as input data for the eating analysis model.

[0056] In an embodiment of the present invention, the second food data is the food information of the patient. When there are missing food ingredients in the meal plan, there will be a situation where the food eaten is inconsistent with the meal plan, such as the fruit in the meal plan is apple, but the fruit eaten on that day is pear; the dish in the meal plan is steamed fish, but the actual dish eaten is spicy fish. The preset threshold is adjusted according to the patient's personal information and eating ability. For example, apples and pears are fruits that patients need to eat on a daily basis, so the judgment threshold of this type of food is set lower, that is, the similarity value judgment threshold of the fruit category is lower; and if the doctor's order stipulates that the patient should not eat spicy food, the preset threshold of the dish is set higher, and the similarity value judgment between steamed fish and spicy fish is lower than the preset threshold.

[0057] Furthermore, the system also includes a prompt module, which is used to send prompt information to the user.

[0058] When it is determined that the similarity value between the first food data and the second food data is lower than a preset threshold, the prompt module sends a prompt message to the user, wherein the prompt message includes prompting the user to verify whether the food that the patient currently needs to eat meets the eating requirements.

[0059] When the information input by the user through the input collection unit meets the eating requirements, the control module executes the operation of allowing the patient to eat, and uses the second food data as input data of the eating analysis model.

[0060] For example, the dish in the meal plan is steamed fish (e.g., with a slight spiciness), but the actual dish is identified as spicy fish. At this time, the similarity value is judged to be lower than the preset threshold. However, the patient's dietary taboos and medical advice do not specify the spiciness, so the patient can enter the food that meets the requirements. At this time, the second food data, spicy fish, is used as the input data of the meal analysis model.

[0061] Through the above operation, the food actually eaten by the patient can be made as close as possible to the eating plan, ensuring the nutritional balance of the food consumed by the patient, and preventing the patient from eating food related to food taboos and endangering personal safety. In addition, the accuracy of the input data of the eating analysis model can be improved, thereby improving the accuracy of the analysis results of the patient's eating ability.

[0062] Furthermore, in an embodiment of the present invention, a fixing module is included that can fix tableware, such as a transparent lunch box with a locking function, and the control module locks and opens the lunch box to determine whether eating is allowed.

[0063] Furthermore, the sensor unit also includes a temperature detector, which detects the temperature of the food to be eaten by the patient.

[0064] The food temperature is obtained before the operation of allowing the patient to eat is performed, and the operation of allowing the patient to eat is performed when it is determined that the food temperature is lower than the preset food eating temperature.

[0065] In the embodiment of the present invention, the above temperature determination operation can prevent patients from being scalded by over-scalding food.

[0066] Furthermore, the patient's eating data is data collected in real time. The data analysis unit obtains the patient's eating data in real time, and performs real-time analysis based on the patient's eating data to determine whether the current eating data triggers an alarm event. When an alarm event is triggered, an alarm is issued to the user through the prompt module.

[0067] Specifically, in an embodiment of the present invention, the patient's eating data can be analyzed by the data analysis unit. For example, if the weight of the tableware changes too quickly and triggers an alarm event, the prompt module will issue an alarm prompt to the user to verify whether the patient has other situations such as swallowing too quickly or spilling the meal.

[0068] Through the above operations, the patient's eating process can be further monitored to ensure the patient's safety.

[0069] Furthermore, the specific operation of obtaining the eating analysis model through training based on the machine learning algorithm by the model training unit is:

[0070] Collect patient personal information and first eating data of several users, including but not limited to patient age, gender, food taste, food taboos, chewing ability, name of food eaten, name of dish eaten, eating time, eating speed, and eating degree.

[0071] The above patient personal information and the first eating data are used as input data for the training model, the eating ability score and the preference value are used as model output data, and the model is trained based on a machine learning algorithm to obtain an eating analysis model.

[0072] Specifically, the embodiment of the present invention collects the patient personal information and the first eating data of several users as the training data set when training the model, with the purpose of building an eating analysis model with a benchmark judgment value. Since the input data is the patient's data, the eating ability score is used as the output data of the model. In addition, the patient's own preference value for the food or dish eaten is also used as the output data, so as to establish a connection between the input data such as the eating degree, eating speed, and food taste and the preference value.

[0073] After obtaining the patient's eating ability and preference value through the eating analysis model, the user can adjust the eating plan based on the analysis results. For example, when different foods have the same nutritional value, choose the food with a higher patient preference value to further improve the patient's eating enthusiasm.

[0074] Furthermore, the data storage module also stores the patient's eating analysis results, and the data processing and analysis module can count the changes in the patient's eating habits over a period of time and transmit them to the display module in the form of charts for user reference.

[0075] Furthermore, the machine learning algorithm includes a random forest algorithm.

[0076] Specifically, in the embodiment of the present invention, the specific method of constructing the eating analysis model by the random forest algorithm is:

[0077] Define the input matrix X∈R n×m , where n is the number of samples and m is the number of features. The features include but are not limited to the patient's age, gender, food taste, food taboos, chewing ability, name of food eaten, name of dish eaten, eating time, eating speed, and eating degree.

[0078] Define the output matrix as Y∈R n×2 , where the first column Y1 is the eating ability score and Y2 is the preference value.

[0079] For the i-th decision tree t i , the training subset is (X i , Y i ), where X i and Y i It is a subset randomly selected from the original data with replacement. The splitting basis of the decision tree is to maximize the information gain:

[0080]

[0081] Among them, f i is a feature, p k is the sample distribution probability, ΔH(Y k ) is the uncertainty reduction of the label before and after the split, that is, the information gain, K is the number of features, It means traversing and calculating all possible split points k.

[0082] Maximize the information gain G(f i ) measures the selection feature f i The reduction in information uncertainty brought about by splitting. At each node, the decision tree selects the feature f that maximizes the information gain. i When building a decision tree, the best features are selected for splitting, ensuring that each split can minimize uncertainty and make the samples in each child node as consistent as possible in the output label, thereby improving the prediction ability of the decision tree.

[0083] Each decision tree predicts the input feature matrix X and outputs two values: eating ability score and preference value.i The predicted output is:

[0084]

[0085] in, is the prediction of the eating ability score, is the prediction of the like value.

[0086] The final output of the random forest model is the weighted average of the prediction results of all decision trees. The random forest consists of T decision trees. Assume that the weight of each decision tree is w i , the final output is:

[0087]

[0088] Where T is the number of decision trees, w i is the weight of each decision tree, t i (X)1 is the prediction of the eating ability score by the ith decision tree, t i (X)2 is the prediction of the preference value by the i-th decision tree.

[0089] Finally, the data, eating plan, eating data, food data, etc. collected by the data acquisition module recorded in the eating ability monitoring system for elderly dementia patients based on machine learning in this embodiment are only preferred embodiments of the present invention, and are only used to illustrate the technical scheme of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical schemes recorded in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. These modifications or replacements do not deviate the essence of the corresponding technical schemes from the spirit and scope of the technical schemes of the various embodiments of the present invention.

Claims

1. A system for monitoring the eating ability of patients with Alzheimer's disease based on machine learning, characterized in that: The system includes a data acquisition module, a data transmission module, a data storage module, a data processing and analysis module, and a display module; The data collection module is used to collect patient personal information, patient food data, and patient eating data; The data transmission module is used to transmit the data collected by the data collection module to the data storage module; The data storage module is used to store the data collected by the data collection module; The data processing and analysis module includes a data processing unit, a model training unit and a data analysis unit; the data processing unit is used to retrieve the data collected by the data collection module from the data storage module for data processing to obtain processed data; The data analysis unit is used to input the processed data into the eating analysis model, and output the analysis results through the eating analysis model; the eating analysis model is obtained by training the model training unit based on the machine learning algorithm; The data transmission module is also used to transmit the output information of the data processing and analysis module to the display module, and the user adjusts the patient's eating plan according to the analysis results of the model training unit.

2. The system for monitoring the eating ability of patients with Alzheimer's disease based on machine learning according to claim 1 is characterized in that: The data acquisition module includes an input acquisition unit, a video monitoring unit and a sensor unit; The input collection unit is used to provide an input terminal for the user to input information, and collect the information input by the user, including collecting the patient's personal information and the food information of the patient's eating plan input by the user; wherein the food information in the eating plan input by the user is used as the first food data; The video monitoring unit is used to collect food data of the patient's eating plan and the patient's eating data during the eating process; the food image data in the eating plan collected by the video monitoring unit is used as the second food data, and the patient's eating plan food data includes the first food data and the second food data; the food image of the actual eating collected by the video monitoring unit is used as the first eating data; The sensor unit is used to collect the patient's actual eating state during the eating process through the sensor and use it as the second eating data. The patient's actual eating data includes the first eating data and the second eating data.

3. The system for monitoring the eating ability of patients with Alzheimer's disease based on machine learning according to claim 2 is characterized in that: The data processing unit performs data processing on the data collected by the data collection module, including performing image recognition processing on the second food data and the first eating data respectively.

4. The system for monitoring the eating ability of patients with Alzheimer's disease based on machine learning according to claim 2 is characterized in that: The patient's personal information includes the patient's age, gender, food taste, food taboos, chewing ability, and disease diagnosis history; The first food data includes food names and dish names input by the user according to the patient's eating plan; The second food data includes food name and dish name; The first eating data includes the name of the food and dish eaten, eating time, eating speed, and eating degree; The second eating data includes the tilt angle and weight of the tableware.

5. The system for monitoring the eating ability of patients with Alzheimer's disease based on machine learning according to claim 3 is characterized in that: The system also includes a control module; Before the patient eats, the data analysis unit determines the similarity value between the first food data and the second food data. When it is determined that the similarity value between the first food data and the second food data is greater than or equal to a preset threshold, the control module executes an operation to allow the patient to eat, and uses the second food data as input data for an eating analysis model.

6. The system for monitoring the eating ability of patients with Alzheimer's disease based on machine learning according to claim 5 is characterized in that: The system also includes a prompt module, which is used to send prompt information to the user; When it is determined that the similarity value between the first food data and the second food data is lower than a preset threshold, the prompt module sends a prompt message to the user, wherein the prompt message includes prompting the user to verify whether the food that the patient currently needs to eat meets the eating requirements.

7. The system for monitoring the eating ability of patients with Alzheimer's disease based on machine learning according to claim 6 is characterized in that: The sensor unit includes a temperature detector, through which the temperature of the food to be eaten by the patient is detected; The food temperature is obtained before the operation of allowing the patient to eat is performed, and the operation of allowing the patient to eat is performed when it is determined that the food temperature is lower than the preset food eating temperature.

8. The system for monitoring the eating ability of patients with Alzheimer's disease based on machine learning according to claim 7 is characterized in that: The patient's eating data is data collected in real time. The data analysis unit obtains the patient's eating data in real time, and performs real-time analysis based on the patient's eating data to determine whether the current eating data triggers an alarm event. When an alarm event is triggered, an alarm is issued to the user through the prompt module.

9. The system for monitoring the eating ability of patients with Alzheimer's disease based on machine learning according to any one of claims 1 to 8, characterized in that: The input end of the model training unit receives the processed data and outputs the analysis results through the eating analysis model. Specifically, the processed data is the patient's personal information and the first eating data, and the analysis results include the eating ability score and the preference value.

10. The machine learning-based eating ability monitoring system for patients with Alzheimer's disease according to claim 9, characterized in that: The machine learning algorithm includes a random forest algorithm. Each decision tree in the random forest model outputs a predicted value of each processed data for the eating ability score and the preference value, and the final output is a weighted average of the prediction results of all decision trees.