Nutrition and health assessment method and system based on big data

Abstract

The invention relates to the technical field of health assessment, in particular to a nutrition health assessment method and system based on big data, and the method comprises the following steps: obtaining diet time and dividing time periods, analyzing nutrition continuous intake behaviors, matching physical sign data to recognize fluctuation events, extracting associated components, constructing an assessment matrix, and outputting a health report. According to the method, food frequency density summarization is introduced, dynamic period coverage analysis of nutrition intake data is achieved, linkage monitoring of a nutrition metabolism time window and health sign changes is combined, and a cross feature system of continuous intake behaviors and physiological fluctuation is constructed; a three-dimensional matrix structure is synchronously introduced to carry out time, structure and response multi-dimensional aggregation evaluation, multi-angle quantitative analysis of the causal relationship between individual nutrition behaviors and health states is realized, the individual health conditions are truly reflected, the timeliness, pertinence and constitutive property of nutrition health evaluation are improved, and the evaluation accuracy is improved. And the sensitivity of health risk identification and the suitability of intervention suggestions are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of health assessment technology, and in particular to a nutrition and health assessment method and system based on big data. Background Art

[0002] The field of health assessment technology involves various methods and means for quantitative and qualitative analysis of individual health status, including the collection and management of physiological indicator monitoring data, the construction and optimization of disease risk prediction models, the formulation of health intervention strategies and their dynamic evaluation, etc. It covers the entire process from data acquisition, analysis and processing to evaluation feedback, and is an important supporting technology system in scenarios such as modern medical services, public health management, and personalized health guidance. Among them, the traditional nutritional health assessment method refers to manually recording dietary information through questionnaires and combining it with static analysis of basic physiological indicators obtained from physical examination items to determine the match between individual nutritional intake and health status. It usually uses regular manual interviews with paper record sheets, nutritional calculation formulas, and manual entry of dietary guidelines into the database to achieve data collection and preliminary analysis, and to evaluate the complex relationship between individual health status and nutritional intake structure in real time.

[0003] The traditional nutritional health assessment process mainly relies on manual questionnaires and static physical examination indicators for analysis. Under the conditions of insufficient continuity and lack of timeliness, it is difficult to reflect the dynamic relationship between an individual's actual nutritional intake rhythm and changes in physiological responses. Especially when the intake interval is short or the intake of complex nutrients is combined, the lack of timestamp records and continuous tracking methods can easily lead to data fragmentation and distortion of analysis results. For example, in scenarios where multiple intakes lead to accumulated fluctuations in physical signs, it is impossible to accurately identify the influence chain of nutritional factors, resulting in a lack of targetedness and early warning basis in the formulation of nutritional intervention measures. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the existing technology and propose a nutrition and health assessment method and system based on big data.

[0005] In order to achieve the above objectives, the present invention adopts the following technical solution: a nutrition and health assessment method based on big data, comprising the following steps:

[0006] S1: Obtain the intake time from the individual dietary record data, set the criteria for dividing the dietary period, compare the intake time with the boundaries of each time period item by item, summarize the frequency density of food intake per unit time period, and generate the food intake period coverage;

[0007] S2: Based on the segmented dietary items in the food intake period coverage, extract the corresponding nutrient component details and intake timestamps, segment the continuous sequence by comparing the nutrient metabolism time window threshold, and generate a continuous nutrient intake behavior feature set;

[0008] S3: Reading the continuous time period of each component intake in the continuous nutritional intake behavior feature set, matching the health sign data change records within the coverage interval, and using the fluctuation threshold to judge the difference between adjacent records item by item and identify fluctuation events, thereby generating an analysis result of the relationship between nutritional intake and health fluctuations;

[0009] S4: Based on the analysis results of the relationship between nutritional intake and health fluctuations, analyzing the overlap in fluctuation events, counting the nutrients that appear within the threshold of significant correlation events, and generating a list of nutrients associated with health status fluctuations;

[0010] S5: Based on the list of nutrients associated with health status fluctuations, combined with the intake time period, continuous intake behavior of ingredients, and fluctuation events, a three-dimensional matrix for nutrition and health assessment is established to generate a comprehensive nutrition and health assessment report.

[0011] As a further solution of the present invention, the food intake time period coverage includes the time period label density distribution, the food intake coverage ratio, and the number of food types consumed in a unit time period; the continuous nutritional intake behavior feature set includes the duration of nutritional components, the continuous intake frequency, and the nutrient component event sorting sequence; the nutritional intake and health fluctuation relationship analysis results include the correspondence between intake components and physical sign fluctuations, the number of physical sign fluctuation events, and the intake behavior correlation strength; the health status fluctuation-associated nutritional component list includes significant nutritional component items, duration records of each component, and statistics of the number of overlapping events; the comprehensive nutritional health assessment report includes the time distribution dimension structure, nutritional balance dimension indicators, and physiological response dimension parameters.

[0012] As a further solution of the present invention, the specific steps for obtaining the food intake period coverage are:

[0013] S111: Obtaining the intake time in the individual's dietary record data, comparing it with the set time segment boundary, assigning a label corresponding to the corresponding time segment, and generating intake time segment label data;

[0014] S112: Based on the intake time period label data, all records are classified and aggregated according to the time period, the number of records in which each food name appears in each time period is counted, and the frequency density value of food intake per unit time period is calculated to generate food intake frequency density data per unit time period;

[0015] S113: Mark each food according to the frequency of occurrence of each food in each time period in the food unit time period intake frequency density data, and summarize the recorded time periods of each food in the five time periods to generate food intake time period coverage.

[0016] As a further solution of the present invention, the intake frequency density value per unit time period adopts the formula:

[0017] ;

[0018] Calculate, where: Indicates the first Food in the The frequency density of the intake in the time period (unit: items / minute). Indicates the The number of records of this type of food in this time period (unit: record), Indicates the length of the time period (in minutes).

[0019] As a further embodiment of the present invention, the specific steps for obtaining the continuous nutrition intake behavior feature set are:

[0020] S211: Based on the food intake period coverage, extract the nutrient name and intake timestamp corresponding to each record, sort all records of the same nutrient in ascending order according to the time value of the timestamp, calculate the time interval between two adjacent records, and compare the time interval value with the nutrient metabolism time window threshold to determine whether the adjacent records belong to a continuous event, thereby generating a nutrient continuous event marker set;

[0021] S212: Based on the record sequence marked as continuous in the nutrition continuous event mark set, each group of continuous records is grouped by ingredient name and the corresponding duration is calculated, and the record group with a duration greater than the effective intake duration threshold is selected and the corresponding nutrition ingredient label is retained to generate effective nutrition duration data;

[0022] S213: According to the duration of each type of nutrient in the effective nutrition duration data, the occurrence frequency records of each type of nutrient are counted, and an attribute set indexed by the nutrient is constructed to establish a continuous nutrition intake behavior feature set.

[0023] As a further embodiment of the present invention, the specific steps for obtaining the analysis results of the relationship between nutritional intake and health fluctuations are as follows:

[0024] S311: Acquire individual health sign data, extract measurement timestamps and corresponding values, combine the corresponding continuous intake time periods of each nutrient component in the continuous nutrition intake behavior feature set, match the health sign data covering the corresponding time period by component, and generate a sequence of corresponding health sign values within the time period;

[0025] S312: Based on the continuous records of each sign in the sequence of values corresponding to the health signs within the time period, the differences between the values of adjacent records are calculated in sequence, and the differences are compared with the fluctuation thresholds respectively. All records that meet any threshold condition are marked to generate a health sign fluctuation event label set;

[0026] S313: Based on the event tags identified in the health sign fluctuation event tag set, combine and merge them according to nutritional components and physical sign types, extract the number of fluctuation events in each combination and record the nutritional component index corresponding to the distribution, and establish a numerical result of the relationship between nutritional intake and health fluctuations.

[0027] As a further solution of the present invention, the specific steps for obtaining the list of nutrients associated with health status fluctuations are:

[0028] S411: Based on the numerical results of the relationship between nutrient intake and health fluctuations, extract the time of fluctuation events and the continuous intake intervals of the nutrient components, compare the time points of each fluctuation event in chronological order to see if they fall into the time intervals of different nutrient components, determine whether the time points of fluctuation events corresponding to multiple nutrients overlap, and generate a record of health fluctuation time overlaps;

[0029] S412: Based on the event records marked as time overlap in the health fluctuation time overlap records, group the occurrences of the nutrients by name and count them, retain the nutrient records with a frequency of occurrence that meets the threshold of significant associated events, and extract the corresponding intake duration to generate repeated verification stable intake data;

[0030] S413: Based on the repetition times and duration values of each nutrient in the repeated verification stable intake data, a data list format is constructed with the nutrient name as the main index and the cumulative intake time as the auxiliary attribute, and various types of ingredients and their intake time are summarized and listed to establish a list of nutrients associated with health status fluctuations.

[0031] As a further solution of the present invention, the specific steps for obtaining the comprehensive nutrition and health assessment report are as follows:

[0032] S511: Based on the index of each ingredient in the health status fluctuation-associated nutrient list, combined with the corresponding intake time segment and continuous intake time length data, a daily intake time distribution table is constructed by ingredient name, and the coverage quantity of each ingredient in the five time periods is counted to establish a nutrient time distribution record;

[0033] S512: Based on the coverage amount of all ingredients in the nutrient component time distribution record in each time period, combined with the daily intake duration and intake frequency, calculate the balance coefficient of each ingredient intake behavior, and summarize and establish a nutrient intake balance index;

[0034] S513: Based on the nutritional intake balance index and the number and duration of physical sign fluctuations recorded in the nutritional component list associated with the health status fluctuation, a three-dimensional data matrix indexed by components is constructed to generate a comprehensive nutritional health assessment report.

[0035] As a further solution of the present invention, the equalization coefficient adopts the formula:

[0036] ;

[0037] Calculate, where Indicates the The balance coefficient of intake behavior of various nutrients, Indicates the The nutrients in The number of coverages within a time period, Indicates the The frequency of intake of the nutrients during the statistical period, Indicates the The total duration of intake of the nutrient component during the statistical period (in hours), represents the average duration of intake of all nutrients (in hours), represents the average frequency of intake of all nutrients (dimensionless), Indicates the absolute value, represents the square root operation, Represents the summation operation for the five intake time periods (early morning, morning, afternoon, evening, and night).

[0038] A nutrition and health assessment system based on big data, including:

[0039] The dietary segmentation module obtains individual dietary intake time, sets the dietary period division standard, compares the time to determine the segment label, calculates the intake frequency density per unit time period in each segment, and generates the food intake period coverage;

[0040] The component sequence extraction module extracts the nutritional components and timestamps according to the food intake period coverage, determines the relationship between adjacent intervals and the nutritional metabolism time window threshold, summarizes the continuous intake information, and generates a continuous nutritional intake behavior feature set;

[0041] The physical sign fluctuation matching module extracts the continuous time period of the components based on the continuous nutritional intake behavior feature set, obtains the health physical sign data records within the covered period, determines whether the difference between adjacent records exceeds the fluctuation threshold, determines the corresponding relationship between intake and fluctuation events, and generates an analysis result of the relationship between nutritional intake and health fluctuation;

[0042] The abnormal component screening module counts the overlapping components based on the analysis results of the relationship between nutritional intake and health fluctuations, screens the items that meet the threshold of significant correlation event repeat verification criteria, and generates a list of nutritional components associated with health status fluctuations;

[0043] The health assessment matrix module is based on the list of nutrients associated with health status fluctuations, extracts time segments, intake behaviors and fluctuation records, constructs a three-dimensional matrix, and generates a comprehensive nutrition and health assessment report.

[0044] Compared with the prior art, the advantages and positive effects of the present invention are:

[0045] In the present invention, by introducing food frequency density aggregation, dynamic time period coverage analysis of nutritional intake data is realized, and the linkage monitoring of nutritional metabolism time window and health sign changes is combined to construct a cross-feature system of continuous intake behavior and physiological fluctuations. Simultaneously, a three-dimensional matrix structure is introduced to perform multi-dimensional aggregation evaluation of time, structure and response, so as to realize multi-angle quantitative analysis of the causal relationship between individual nutritional behavior and health status, truly reflect the individual's health status, improve the timeliness, pertinence and structure of nutritional health assessment, and enhance the sensitivity of health risk identification and the adaptability of intervention recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 Schematic diagram of the steps of the present invention;

[0047] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0049] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0050] See also Figure 1 , a nutrition and health assessment method based on big data, comprising the following steps:

[0051] S1: Obtain the intake time from individual dietary record data, set the dietary time division standard (refer to the five-time division method recommended in the "Dietary Guidelines for Chinese Residents 2022", including 5:00-9:00 in the morning, 9:01-12:00 in the morning, 12:01-15:00 in the afternoon, 15:01-19:00 in the evening, and 19:01-24:00 at night), compare the intake time with the boundaries of each time period item by item, determine the time period to which each record belongs, and mark the time period label simultaneously, summarize the frequency density of food intake per unit time period, and generate food intake time period coverage;

[0052] S2: Based on the segmented dietary items in the food intake period coverage, extract the corresponding nutrient component details and intake timestamps, arrange the occurrence events of each component in chronological order, compare the nutrient metabolism time window threshold (set to 2 hours according to the "Technical Specifications for Dietary Surveys"), determine whether a continuous event sequence is formed, segment the continuous sequence, and record the duration (based on the effective intake duration defined in the "Clinical Nutrition Monitoring Specifications") and frequency of each nutrient component to generate a continuous nutrient intake behavior feature set;

[0053] S3: Obtain individual health sign data, including monitoring items such as blood pressure and heart rate, read the continuous time period of each component intake in the continuous nutrition intake behavior feature set, match the health sign data change records within the coverage interval, and use fluctuation thresholds (for example, blood pressure fluctuation threshold: refer to the "Guidelines for the Prevention and Treatment of Hypertension in China" ≥10mmHg / 10min, heart rate variability threshold: using the SDNN indicator ≥50ms, etc.) to judge the difference between adjacent records item by item, identify fluctuation events, and generate analysis results on the relationship between nutrition intake and health fluctuations;

[0054] S4: Based on the results of the analysis of the relationship between nutritional intake and health fluctuations, analyze the overlap in fluctuation events, count the nutrients that appear within the threshold of significant association events (based on the principle of three-repeat verification recommended in the "Nutritional Epidemiology Research Methods"), and record the corresponding duration to generate a list of nutrients associated with health status fluctuations;

[0055] S5: Based on the list of nutrients associated with health status fluctuations, combined with the corresponding intake time segments, continuous intake behavior of the ingredients, and fluctuation events, a three-dimensional matrix for nutritional health assessment (including: time distribution dimension, nutritional balance dimension, and physiological response dimension) is established to generate a comprehensive nutritional health assessment report.

[0056] The coverage of food intake time periods includes the time period label density distribution, food intake coverage ratio, and the number of food types consumed in a unit time period. The continuous nutritional intake behavior feature set includes the duration of nutrients, continuous intake frequency, and the nutrient event ranking sequence. The analysis results of the relationship between nutritional intake and health fluctuations include the correspondence between intake components and physical sign fluctuations, the number of physical sign fluctuation events, and the strength of the association with intake behavior. The list of nutrients associated with health status fluctuations includes significant nutritional component items, duration records of each component, and statistics of the number of overlapping events. The comprehensive nutritional health assessment report includes the time distribution dimension structure, nutritional balance dimension indicators, and physiological response dimension parameters.

[0057] The specific steps of S1 are:

[0058] S111: Obtain the intake time in the individual dietary record data, compare it with the set time segment boundaries (referring to the five-time division method recommended in the "Dietary Guidelines for Chinese Residents 2022", including the five segments of 5:00-9:00 in the morning, 9:01-12:00 in the morning, 12:01-15:00 in the afternoon, 15:01-19:00 in the evening, and 19:01-24:00 at night)), assign corresponding labels to the corresponding time segments, and generate intake time segment label data;

[0059] To obtain the intake time in individual dietary record data, we should first clarify the fields contained in the original dietary record data, extract the "intake time" field one by one and convert it into a standard time format (such as YYYY-MM-DDHH:MM:SS), and then compare the intake time with the upper and lower boundaries of the five specific time intervals according to the time segments set in the "Dietary Guidelines for Chinese Residents 2022". The comparison operation needs to determine whether the hour and minute values are between the upper and lower limits of each time period one by one. The judgment process extracts the "hour" and "minute" values, converts them into minutes (i.e. hour×60+minute), and then compares them with the start and end boundaries of each time period (such as early morning segment: 5:00 is 3 00 minutes, 9:00 is 540 minutes) to determine the belonging interval. For example, if the intake time is 08:15, the corresponding minute value is 495. Since 495∈[300, 540], it is classified into the 5:00–9:00 morning segment and marked as "early morning". This operation is repeated for all records, and finally an "intake time period label" is added to each record. For example, the intake time of record A is 11:50, which is marked as "morning", record B is 15:05, which is marked as "evening", and record C is 21:30, which is marked as "night". Through this method of item-by-item comparison and segment boundary judgment, the intake time period label assignment operation is completed for all individual dietary records.

[0060] S112: Based on the intake time period label data, all records are classified and aggregated according to the time period, the number of records that appear for each food name in each time period is counted, and the frequency density value of food intake per unit time period is calculated to generate food intake frequency density data per unit time period;

[0061] The frequency density value of intake per unit time period is calculated using the formula:

[0062] ;

[0063] Calculate, where: Indicates the first Food in the The frequency density of the intake in the time period (unit: items / minute). Indicates the The number of records of this type of food in this time period (unit: record), Indicates the length of the time period (in minutes).

[0064] The time period is selected as the morning period. According to the time division standards in the "Dietary Guidelines for Chinese Residents 2022", the morning period is from 9:01 to 12:00, and the time range is 179 to 240 minutes (the number of minutes is hours × 60 + minutes);

[0065] , indicating the starting minute of the morning time period is 9:01 (9×60+1);

[0066] , indicating the end minute of the morning time period is 12:00 (12×60);

[0067] The length of time: ;

[0068] Substituting into the formula:

[0069] ;

[0070] The results show that the frequency density of milk intake per unit time period in the morning time period is This value, as one of the final result data in module S112, will be recorded in the frequency density data table and used for threshold determination and intake time period coverage vector generation in the subsequent step S113. This density value is compared with the frequency existence threshold of 0.005 items / minute. If it is less than the threshold, the intake during this period is marked as invalid, and the coverage status of this food in the morning time period is set to 0.

[0071] The intake frequency density value per unit time period indicates the average occurrence intensity of the intake behavior of a certain type of food in a specific time period on a unit time scale. This value comprehensively reflects the frequency with which the food is recorded or selected by individuals in the time period. The larger the value, the higher the intake concentration of the food in the time period, and the more obvious the time period intake characteristics. This indicator not only takes into account the number of food intake records, but also introduces the degree of deviation of the intake time relative to the center of the time period and the distribution of records of all foods, thereby effectively removing the influence of time range and sample density on frequency statistics, and providing a quantitative basis for further analysis of food intake tendencies, rhythmic patterns and dietary behavior patterns in different time periods throughout the day.

[0072] S113: Mark each food according to the frequency of occurrence of each food in each time period in the food unit time period intake frequency density data, and summarize the recorded time periods of each food in the five time periods to generate food intake time period coverage;

[0073] Based on the aforementioned food intake frequency density data per unit time period, it is necessary to statistically analyze the frequency of each food in the five time periods. First, a two-dimensional table structure is established for the frequency data. The row field is the food name, the column field is the five time period labels, and the cell records the frequency density value of the food in the corresponding time period. The frequency density data of the food in the five time periods are read one by one, and the judgment standard for the existence of frequency is set. For example, when the frequency density is greater than 0.005 / minute, it is determined that there is an intake record in the time period. The threshold value can be adjusted based on the average value of all frequency densities. For example, if the frequency density mean in the total sample is 0.012, the set 0.005 can be regarded as the lower limit value. The corresponding logical judgment is: if the frequency density of a certain food in the "morning" period is , does not meet the existence condition, marked as 0; if it is 0.00628, it meets the condition and is marked as 1. In this way, the five time periods are judged one by one to form a coverage vector similar to [0, 0, 1, 1, 0], and then the intake time coverage of the food throughout the day is summarized. For example, if the frequency of "bread" in the early morning, morning and night meets the threshold, then its coverage time period is "early morning, morning, night", and it is recorded and stored. Finally, each food corresponds to a coverage time period information, such as "milk-coverage time period: early morning, morning, afternoon, evening, night". If a food has no record with a frequency density greater than the threshold in all time periods, its coverage time period is empty.

[0074] The specific steps of S2 are:

[0075] S211: Based on the food intake period coverage, extract the nutrient name and intake timestamp corresponding to each record, sort all records of the same nutrient in ascending order according to the time value of the timestamp, calculate the time interval between two adjacent records, and compare the time interval value with the nutrient metabolism time window threshold (2 hours) to determine whether the adjacent records belong to a continuous event, thereby generating a nutrient continuous event marker set;

[0076] Based on the food intake period coverage table, extract the nutrient name and intake timestamp field corresponding to each dietary record one by one. The intake timestamp needs to be uniformly converted into a standard time format and further converted into a Unix timestamp (in seconds) to facilitate time difference calculation. All records with the same nutrient name are summarized and sorted in ascending order according to the intake timestamp to ensure that they are arranged in order of time. Then, the time difference is calculated for each group of adjacent records one by one. The calculation method is: subtract the Unix timestamp of the current record from the Unix timestamp of the previous record, and then divide the obtained time difference by 3600 to convert it to hours. For example, the timestamp of record A is 1622505600, and that of record B is 1622512800. The adjacent time interval is (1622512800-1622505600) / 3600=2 Hours, then each group of time interval values is compared with the nutritional metabolism time window threshold of 2 hours. If the time interval is less than or equal to 2 hours, the record pair is determined to be a continuous nutritional intake event and is assigned a "continuous" mark. If it is greater than 2 hours, it is marked as "discontinuous". The entire judgment process is executed in a loop for the records corresponding to each nutrient component, and finally a nutritional continuous event mark is generated for each record, forming a set of nutritional continuous event mark sets. For example, in the four records of vitamin C, the time interval of the first and second is 1.5 hours and is marked as "continuous", the time interval of the second and third is 2.8 hours and is marked as "discontinuous", and the time interval of the third and fourth is 1 hour and is marked as "continuous". The final structure can be expressed as vitamin C: [1, 0, 1], which represents the continuity status between the 1st-2nd, 2nd-3rd, and 3rd-4th records respectively.

[0077] S212: Based on the record sequences marked as continuous in the nutrition continuous event tag set, each group of continuous records is grouped by ingredient name and the corresponding duration is calculated. The record groups that are greater than the effective intake duration threshold are screened and the corresponding nutrition ingredient labels are retained to generate effective nutrition duration data;

[0078] Based on the above-formed nutrition continuous event tag set, the record pairs marked as "continuous" are processed and grouped into a continuous group according to the name of the nutrient component. The records in the group need to be sorted in ascending order by timestamp. The duration of each group of continuous records is further calculated, that is, the timestamp of the latest record in the group is subtracted from the timestamp of the earliest record and converted into hours. For example, a group of calcium records are at 8:00, 9:00 and 10:30 respectively, and their Unix timestamps are 1622505600, 1622509200 and 1622514600 respectively. The duration of this group is (1622514600-1622505600) / 3600=2.5 Hours, after the calculation of each group is completed, the duration is compared with the effective intake duration threshold, and the effective threshold is set to 2 hours. If the duration of a group is greater than 2 hours, the group is retained and the corresponding nutrient label is extracted. If the duration is less than or equal to 2 hours, the group is discarded. After completing the judgment of all record groups, a set of nutrients that only contain effective intake duration records is formed. At the same time, the duration information corresponding to each nutrient is retained in the output. For example, the effective calcium intake group lasts for 2.5 hours and is retained, and a group of iron lasts only 1.2 hours and is discarded. The final structure is generated as "nutrient: calcium, duration: 2.5 hours", and all nutrient items that meet the effective duration conditions are retained.

[0079] S213: Based on the duration of each type of nutrient in the effective nutrition duration data, count the occurrence frequency records of each type of nutrient, combine and construct an attribute set indexed by the nutrient, and establish a continuous nutrition intake behavior feature set;

[0080] Based on the effective nutrition duration data, each nutrient component is counted item by item. First, the nutrient component labels in all valid records are extracted and counted. The frequency is accumulated every time they appear. For example, if calcium appears in 3 groups, iron appears in 1 group, and vitamin C appears in 2 groups, the frequency of calcium is 3, iron is 1, and vitamin C is 2. These frequencies are combined with the duration of each group to construct an attribute set. The attribute fields include "nutrient name", "occurrence frequency", "average duration", "maximum duration", and "minimum duration". For example, if the duration of three valid records of a nutrient component calcium is 2.5 hours, 3 hours, and 2.8 hours respectively, the average duration is (2.5+3+2.8) / 3=2.77 hours, with a maximum of 3 hours and a minimum of 2.5 hours. Finally, a continuous nutrition intake behavior feature set is generated with the nutrient component as the index item. Each nutrient component corresponds to a complete set of behavioral feature records, such as "calcium: frequency 3 times, average duration 2.77 hours, maximum 3 hours, minimum 2.5 hours". Based on this, a complete structured feature set dataset is formed.

[0081] The specific steps of S3 are:

[0082] S311: Obtain individual health sign data (blood pressure, heart rate data), extract measurement timestamps and corresponding values, combine the continuous nutritional intake behavior feature set with the corresponding continuous intake time period of each nutrient component, match the health sign data covering the corresponding time period by component, and generate a sequence of health sign values corresponding to the time period;

[0083] After obtaining the original records of blood pressure and heart rate in the individual health sign data, it is necessary to extract the measurement timestamp field and the corresponding sign value field one by one. The blood pressure data appears as a pair of systolic pressure and diastolic pressure, and two values are extracted respectively. The heart rate is calculated based on the heartbeat interval to obtain the SDNN value, and the unit is ms; the extracted timestamp must be in a unified format of YYYY-MM-DDHH:MM:SS and converted into a Unix timestamp to facilitate the subsequent overlap determination with the nutritional intake time period; combined with the intake time period information corresponding to each nutrient component in the continuous nutritional intake behavior feature set obtained above, traverse each nutrient component and extract its "intake start timestamp" and "intake end timestamp" recorded in the valid intake record, and filter the health sign data within this time period. The screening method is to determine whether the timestamp of the health sign record meets the following requirements: start timestamp ≤ physical sign timestamp ≤ end timestamp. If it meets the requirements, Then extract the record, and the extracted content is the combination of the corresponding vital sign value and the timestamp. Repeat the judgment process to traverse all nutrients and their intake time periods, and finally generate a healthy vital sign value sequence set for each nutrient, where each record corresponds to a blood pressure or heart rate value and a timestamp in the time period. For example, if the vitamin D intake time period is 10:00-13:00, if the blood pressure records appear at 10:15, 11:45, and 12:30 respectively, and the measured values are 126 / 84, 118 / 78, and 122 / 80, then the sequence is {10:15-126 / 84, 11:45-118 / 78, 12:30-122 / 80}. If the heart rate record is {10:20-SDNN=42ms, 12:10-SDNN=56ms}, then the vital sign value sequence corresponding to the nutrient is established by type to establish independent lists of blood pressure and heart rate and output.

[0084] S312: Based on the continuous records of each sign in the corresponding value sequence of the health sign within the time period, the difference between the values of adjacent records is calculated in sequence, and the difference is compared with the fluctuation threshold (blood pressure fluctuation threshold 10mmHg, heart rate SDNN threshold 50ms). All records that meet any threshold condition are marked to generate a health sign fluctuation event label set;

[0085] For the sequence of vital sign values obtained under each nutrient component, the blood pressure and heart rate records are processed separately. When processing blood pressure, all systolic blood pressure (SBP) and diastolic blood pressure (DBP) values in the sequence are extracted, arranged in chronological order, and the differences between adjacent records are calculated in sequence. The calculation method is to subtract the previous record from the current record. For example, if the sequence is [126, 118, 122], the differences are -8 (118–126) and +4 (122–118), respectively. The absolute value of each calculated difference is taken and compared with the blood pressure fluctuation threshold of 10 mmHg. If the absolute value is ≥10, the record pair is identified as a blood pressure fluctuation event. The heart rate processing is similar. The SDNN value sequence is extracted, and the differences between adjacent records are calculated in sequence and the absolute value is taken. The value is compared with the SDNN fluctuation threshold of 50ms. If the difference is ≥50ms, it is marked as a heart rate fluctuation event. For example, if the sequence is [42, 56, 49], the difference is +14 and -7, both less than 50, and they are not marked. If it is [42, 102], the difference is +60, because 60>50, it is marked. Finally, all records that meet the above threshold judgment conditions are marked to form a health sign fluctuation event label set. The label content includes the start timestamp, end timestamp, sign type, fluctuation value, and whether the fluctuation event is triggered (yes / no). For example, "blood pressure -118→129-difference 11mmHg-yes", "heart rate -42→102-difference 60ms-yes" and other structured label sets.

[0086] S313: Based on the identified event tags in the health sign fluctuation event tag set, group and merge them by nutrient component and physical sign type, extract the number of fluctuation events in each combination and record the nutrient component index corresponding to the distribution, and establish a numerical result of the relationship between nutrient intake and health fluctuation;

[0087] The records in the healthy vital sign fluctuation event tag set are processed and combined and merged according to the two dimensions of nutrient components and vital sign types, such as the nutrient components calcium and blood pressure combination, and calcium and heart rate combination. A combination key value is constructed, and the number of fluctuation events marked as "yes" in each combination is counted. The counting method is to traverse all tags under the corresponding combination and count the number of entries with "triggered fluctuation events" as "yes". For example, there are 5 records under calcium-blood pressure, of which 3 trigger fluctuation events, then the number of records is 3. The number of fluctuations in all combinations is counted in the same way, and the nutrient component index is retained in the result to form a relationship mapping, such as key: "calcium-blood pressure", value: "number of fluctuation events = 3". Finally, a triple set with the structure of {nutrient component name, vital sign type, number of fluctuation events} is constructed, such as {vitamin C, heart rate, 2}, {iron, blood pressure, 0}, {potassium, blood pressure, 4}, etc., completing the summary representation of the numerical results of the relationship between nutrient intake and healthy vital sign fluctuations.

[0088] The specific steps of S4 are:

[0089] S411: Based on the numerical results of the relationship between nutrient intake and health fluctuations, the time of fluctuation events and the continuous intake intervals of the nutrient components are extracted. The time points of each fluctuation event are compared in chronological order to see if they fall within the time intervals of different nutrient components. It is determined whether the time points of fluctuation events corresponding to multiple nutrients overlap, and a health fluctuation time overlap record is generated.

[0090] According to the generated numerical results of the relationship between nutritional intake and health fluctuations, the "fluctuation event time point" field in each record is extracted. This field is taken from the fluctuation occurrence time point in the corresponding health sign fluctuation event tag set. Generally, the record start time or the fluctuation center timestamp can be used as the identification time point. For example, record A is marked with the start time 1622509200 (i.e. 2021-06-0109:00:00), which is used as the fluctuation event time point. At the same time, the continuous intake interval of each nutrient component is traversed and extracted, including the intake start and end timestamp intervals. Each nutrient component corresponds to multiple continuous intake segments; then each fluctuation event time point is compared one by one with the intake time periods of all nutrients. The judgment logic is: if the fluctuation event timestamp T meets the interval condition of a certain nutrient component intake segment, that is, the intake start time ≤ T≤ If the intake end time is reached, it is determined that the nutrient may be associated with the fluctuation event and marked as overlap. The matching process of T in other nutrient intake time periods is continued. If T meets the conditions in the intervals of multiple components, it is marked that there is a multiple nutrient overlap at this time point, and all matching nutrient names, fluctuation timestamps, and matching segments are recorded to form a "healthy fluctuation time overlap record". Each record contains at least the following fields: fluctuation timestamp, matching nutrient name list, corresponding intake time period identifier, overlap type (single / multiple). For example, if an event T=1622509200 is matched in the vitamin C and calcium intake segments at the same time, the record is: "T=1622509200, overlap component=[vitamin C, calcium], type=multiple". Repeat this process to traverse all fluctuation event times to form a complete overlap record set.

[0091] S412: Based on the event records marked as time overlap in the health fluctuation time overlap records, group the nutrient names and count the number of occurrences. Nutrient records with a frequency of occurrence that exceeds the significant associated event threshold (3 times) are retained, and the corresponding intake duration is extracted to generate repeated verification stable intake data.

[0092] Based on the records of health fluctuation time overlap, all overlap events marked as "multiple" are screened and extracted to construct a record subset. Only fluctuation events involving the overlap of two or more nutrients are retained in this subset. All nutrient name fields are disassembled and expanded into single item records. Then, grouping and counting are performed with the nutrient name as the key. The total number of times each nutrient appears in the "multiple time overlap events" is counted to determine whether it exceeds the significant frequency threshold. The threshold is set to 3 times, that is, only nutrient records with a number of occurrences ≥ 3 are retained. For these nutrients that meet the conditions, the length of the intake time period associated with the overlap event is extracted, that is, the length of the intake time period of the ingredient in the overlap. The difference between the start and end time of intake in the overlapping events was converted into hours ((end timestamp - start timestamp) / 3600). The duration of multiple time periods can be accumulated or the maximum and average values can be retained. For example, vitamin D was involved in four overlapping events, and its associated intake periods were 2.5, 3.0, 2.2, and 2.7 hours, respectively. The final average was 2.6 hours, and the maximum was 3.0 hours. The structure fields were "nutrient name, number of occurrences, total intake duration, maximum value, average value", and records such as "vitamin D - 4 occurrences - total duration 10.4h - maximum 3.0h - average 2.6h" constituted a repeated verification stable intake data set.

[0093] S413: Based on the repetition times and duration values of each nutrient in the repeatedly verified stable intake data, a data list format is constructed with the nutrient name as the primary index and the cumulative intake duration as the auxiliary attribute. Each nutrient and its intake duration are summarized and listed to establish a list of nutrients associated with health status fluctuations;

[0094] Based on the repeatedly verified stable intake data set, data records are established for each type of nutrient in turn, with the nutrient name as the primary index and the intake time indicator as the auxiliary attribute. The specific fields include: cumulative intake time (i.e. the total duration of the associated intake time period in all overlapping events, in hours), frequency of overlap (i.e. the number of times participating in "multiple overlap" events), and optional extended fields include the maximum and average intake time. Finally, a data list structure is constructed, with each row recording a type of nutrient, such as "Vitamin C: cumulative intake time of 7.5 hours, 3 overlaps", "Calcium: cumulative 9.0 hours, 4 overlaps". The data list is presented in a structured table format to facilitate subsequent analysis of the statistical correlation between nutrient intake and health fluctuations. Finally, a "nutrient list associated with health status fluctuations" is formed, with each type of ingredient listing its corresponding intake time and overlap intensity value, realizing the structural transformation from overlapping records to classified lists.

[0095] The specific steps of S5 are:

[0096] S511: Based on the index of each ingredient in the health status fluctuation-related nutrient list, combined with the corresponding intake time segment and continuous intake time length data, a daily intake time distribution table is constructed by ingredient name, and the coverage amount of each ingredient in the five time periods is counted to establish a nutrient time distribution record;

[0097] Based on the name of each nutrient in the list of nutrients associated with health status fluctuations as the primary index, combined with each previously recorded intake time segment and corresponding duration data, divided by day and archived one by one, a daily intake behavior record table was established, and the start and end time of the intake time period was extracted in each record. The intersection relationship between this time period and the five time periods defined in the "Dietary Guidelines for Chinese Residents 2022" (5:00-9:00 in the morning, 9:01-12:00 in the morning, 12:01-15:00 in the afternoon, 15:01-19:00 in the evening, and 19:01-24:00 at night) was calculated to determine whether the intersection exists. If it exists, it is judged to be Defined as "coverage", for example, if a certain nutrient is consumed from 08:30 to 11:15 on a certain day, it spans the early morning and morning periods, and should be counted as one coverage in the early morning and morning periods respectively. Each record is judged according to this logic, and the number of times it is covered in the five time periods is counted according to the nutrient. That is, a structured daily intake time distribution table is constructed, with each row corresponding to a nutrient, and each column being the coverage quantity field for five fixed time periods, with the value being the frequency of occurrence of the ingredient in the time period. The final summary record format is such as: "Calcium - early morning: 4, morning: 5, afternoon: 3, evening: 1, night: 0", constructing a complete nutrient time distribution record.

[0098] S512: Based on the coverage of all ingredients in the nutrient component time distribution record in each time period, combined with the daily intake duration and intake frequency, calculate the balance coefficient of each ingredient intake behavior, and summarize and establish the nutrient intake balance index;

[0099] The equalization coefficient is calculated using the formula:

[0100] ;

[0101] Calculate, where Indicates the The balance coefficient of intake behavior of various nutrients, Indicates the The nutrients in The number of coverages within a time period, Indicates the The frequency of intake of the nutrients during the statistical period, Indicates the The total duration of intake of the nutrient component during the statistical period (in hours), represents the average duration of intake of all nutrients (in hours), represents the average frequency of intake of all nutrients (dimensionless), Indicates the absolute value, represents the square root operation, Represents the summation operation for the five intake time periods (early morning, morning, afternoon, evening, and night).

[0102] : refers to the number of times the g-th nutrient component is covered in the h-th time period collected by the continuous intake recording system. The data source is the number of overlaps between the intake time period and the five-stage mapping in the full-day intake record. In the example, the calcium intake record of a person counted by the collection system is: 3 times in the morning, 2 times in the morning, 1 time in the afternoon, 0 times in the evening, and 0 times at night, that is, ;

[0103] : Indicates the frequency of intake of this nutrient on the sampling day, which is automatically counted by the food intake registration frequency statistics system and processed by scanning code or sensor sampling frequency. The system records that calcium is taken in 6 times. ;

[0104] : Refers to the total duration of intake of the nutrient in all intake events on the day. The time period data is sampled and recorded and obtained through the intake behavior tracking terminal. For example, if the record shows 30 minutes of intake in the morning, 45 minutes in the morning, and 20 minutes in the afternoon, a total of 95 minutes, it is uniformly converted to hours. Hour;

[0105] : Refers to the average duration of intake of all nutrients in a day, which is composed of all ingredients The total is divided by the number of ingredients. In this day, there are 10 ingredients and the total intake time is 14.7 hours. Hour;

[0106] : Refers to the average intake frequency of all ingredients. The total sampling frequency of the system is 62 times, and there are 10 types of ingredients. ;

[0107] Substituting the above values into the formula, the step-by-step calculation is as follows:

[0108] Calculate the molecular part

[0109] ;

[0110] calculate:

[0111] ;

[0112] Calculated item by item:

[0113] ;

[0114] ;

[0115] ;

[0116] (twice);

[0117] Numerator total:

[0118] ;

[0119] Calculate the denominator

[0120] ;

[0121] Calculate the sum of squares part:

[0122] ;

[0123] Difference item:

[0124] ;

[0125] ;

[0126] Denominator Total:

[0127] ;

[0128] Obtain the equilibrium coefficient

[0129] ;

[0130] The results indicate a significant bias in the intake of this nutrient across the five time periods, with intake concentrated in the morning and mid-afternoon. The balance coefficient is low, reflecting a significant deviation from the ideal uniform distribution. This value is recorded as an indicator of balanced intake behavior in the subsequent three-dimensional health assessment matrix, along with the time distribution vector and fluctuation intensity parameter, to form the comprehensive assessment attribute for this nutrient. Values closer to 0 indicate a more severe bias in intake, while values closer to 1 indicate a more balanced distribution across the five time periods.

[0131] The balance coefficient indicates the uniformity of the temporal distribution of the intake behavior of a certain nutrient during a day. It is a numerical indicator to measure whether its intake reasonably covers the five key time periods of early morning, morning, afternoon, evening and night. This coefficient comprehensively considers the frequency distribution of the intake of the nutrient in each time period, the overall intake duration and the degree of deviation of the intake frequency in all components. It reflects whether the intake is concentrated in a few time periods or whether it presents a uniform coverage state through relative difference and normalization processing. The closer the coefficient value is to 1, the more uniform the intake distribution, the more reasonable the coverage time period, and the more the intake behavior meets the requirements of balanced nutritional supply throughout the day; the closer the value is to 0, the higher the concentration of intake time is, and there is a risk of intake during off-time periods, which may lead to a decrease in the matching degree of absorption efficiency and physiological rhythm. As an important representation value of nutrient intake balance, this coefficient is directly related to nutritional intervention strategies, time distribution optimization, and the evaluation of time intake regulation related to health fluctuations.

[0132] The formula calculates the distribution deviation value of nutrients in each intake period through the numerator, and uses absolute value operation to measure the difference between the coverage rate under unit frequency and the ideal distribution ratio. The sum is taken for each period and the average is taken to reflect the degree of imbalance of intake behavior on the time axis; the denominator first quantifies the central trend of unit frequency coverage through square summation, and at the same time adds the normalized deviation of intake duration and intake frequency relative to the overall average value. The two are processed by absolute value to ensure positive measurement, and the square root operation is used to standardize the cumulative terms as a whole to suppress the influence of extreme values, forming a balance correction term. The ratio structure of the numerator and denominator is used to comprehensively evaluate the relative relationship between the degree of time coverage deviation and global stability, so that the final balance coefficient reflects both the dispersion of time period distribution and the overall consistency of intake behavior, thereby constructing an intake behavior balance assessment model that emphasizes both quantitative accuracy and practical feasibility.

[0133] S513: Based on the nutritional intake balance index and the number and duration of physiological fluctuations recorded in the nutritional component list associated with health status fluctuations, a three-dimensional data matrix indexed by components is constructed (corresponding to the three indicators of time distribution, intake balance, and physiological fluctuation frequency) to generate a comprehensive nutritional health assessment report;

[0134] The nutritional intake balance index was integrated with the number of physical sign fluctuations and the total duration of intake recorded in the nutrient list associated with health status fluctuations to construct a three-dimensional data matrix indexed by the nutrient component. The three-dimensional indicators in the matrix were time distribution characteristics (represented by a five-time frequency vector, such as [3, 2, 1, 4, 0]), intake behavior balance (balance coefficient E value), and physiological fluctuation intensity (fluctuation intensity S = N / T calculated by the number of physical sign fluctuations N and the total intake duration T; for example, if N = 6 times and T = 12 hours, then S = 0.5 times / hour). These three indicators constitute the multidimensional nutritional and health attributes of the component. Finally, the three-dimensional indicators corresponding to each nutrient component were summarized and formatted into a report. Each record was: "Nutrient: Iron; Time distribution: [2, 2, 1, 0, 0]; Balance coefficient: 0.62; Fluctuation intensity: 0.4 times / hour." All nutrient component records were combined to generate a "Nutrition and Health Comprehensive Assessment Report" to describe the association structure and quantitative characteristics between intake time coverage, behavioral balance, and health fluctuations.

[0135] See also Figure 2 , a nutrition and health assessment system based on big data, including:

[0136] The dietary segmentation module obtains individual dietary intake time, sets the dietary period division standard, compares the time to determine the segment label, calculates the intake frequency density per unit time period in each segment, and generates the food intake period coverage;

[0137] The component sequence extraction module extracts nutrients and timestamps based on the food intake period coverage, determines the relationship between adjacent intervals and the nutrient metabolism time window threshold, summarizes continuous intake information, and generates a continuous nutrient intake behavior feature set;

[0138] The physical sign fluctuation matching module extracts the continuous time period of the components based on the continuous nutritional intake behavior feature set, obtains the health physical sign data records within the covered period, determines whether the difference between adjacent records exceeds the fluctuation threshold, determines the corresponding relationship between intake and fluctuation events, and generates the analysis results of the relationship between nutritional intake and health fluctuations;

[0139] The abnormal component screening module, based on the analysis results of the relationship between nutritional intake and health fluctuations, counts the overlapping components, selects items that meet the threshold of significant correlation events and repeat verification criteria, and generates a list of nutritional components associated with health status fluctuations;

[0140] The health assessment matrix module is based on a list of nutrients associated with health status fluctuations, extracts time segments, intake behaviors and fluctuation records, constructs a three-dimensional matrix, and generates a comprehensive nutrition and health assessment report.

[0141] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A nutritional health assessment method based on big data, characterized in that: The following steps are involved: S1: Obtain the intake time from the individual dietary record data, set the criteria for dividing the dietary period, compare the intake time with the boundaries of each time period item by item, summarize the frequency density of food intake per unit time period, and generate the food intake period coverage; S2: Based on the segmented dietary items in the food intake period coverage, extract the corresponding nutrient component details and intake timestamps, segment the continuous sequence by comparing the nutrient metabolism time window threshold, and generate a continuous nutrient intake behavior feature set; S3: Reading the continuous time period of each component intake in the continuous nutritional intake behavior feature set, matching the health sign data change records within the coverage interval, and using the fluctuation threshold to judge the difference between adjacent records item by item and identify fluctuation events, thereby generating an analysis result of the relationship between nutritional intake and health fluctuations; S4: Based on the analysis results of the relationship between nutritional intake and health fluctuations, analyzing the overlap in fluctuation events, counting the nutrients that appear within the threshold of significant correlation events, and generating a list of nutrients associated with health status fluctuations; S5: Based on the list of nutrients associated with health status fluctuations, combined with the intake time period, continuous intake behavior of ingredients, and fluctuation events, a three-dimensional matrix for nutrition and health assessment is established to generate a comprehensive nutrition and health assessment report.

2. The nutritional health assessment method based on big data according to claim 1, characterized in that: The food intake time period coverage includes the time period label density distribution, food intake coverage ratio, and the number of food types consumed in a unit time period. The continuous nutritional intake behavior feature set includes the duration of nutritional components, continuous intake frequency, and nutritional component event sorting sequence. The nutritional intake and health fluctuation relationship analysis results include the correspondence between intake components and physical sign fluctuations, the number of physical sign fluctuation events, and the intake behavior correlation strength. The list of nutritional components associated with health status fluctuations includes significant nutritional component items, duration records of each component, and statistics of the number of overlapping events. The comprehensive nutritional health assessment report includes the time distribution dimension structure, nutritional balance dimension indicators, and physiological response dimension parameters.

3. The nutritional health assessment method based on big data according to claim 1, characterized in that: The specific steps for obtaining the food intake period coverage are: S111: Obtaining the intake time in the individual's dietary record data, comparing it with the set time segment boundary, assigning a label corresponding to the corresponding time segment, and generating intake time segment label data; S112: Based on the intake time period label data, all records are classified and aggregated according to the time period, the number of records in which each food name appears in each time period is counted, and the frequency density value of food intake per unit time period is calculated to generate food intake frequency density data per unit time period; S113: Mark each food according to the frequency of occurrence of each food in each time period in the food unit time period intake frequency density data, and summarize the recorded time periods of each food in the five time periods to generate food intake time period coverage.

4. The nutritional health assessment method based on big data according to claim 3, characterized in that: The intake frequency density value per unit time period is calculated using the formula: ； Calculate, where: Indicates the first Food in the The intake frequency density value of the time period, Indicates the The number of records of this type of food in this time period, Indicates the length of the time period.

5. The nutritional health assessment method based on big data according to claim 1, characterized in that: The specific steps for obtaining the continuous nutrition intake behavior feature set are: S211: Based on the food intake period coverage, extract the nutrient name and intake timestamp corresponding to each record, sort all records of the same nutrient in ascending order according to the time value of the timestamp, calculate the time interval between two adjacent records, and compare the time interval value with the nutrient metabolism time window threshold to determine whether the adjacent records belong to a continuous event, thereby generating a nutrient continuous event marker set; S212: Based on the record sequence marked as continuous in the nutrition continuous event mark set, each group of continuous records is grouped by ingredient name and the corresponding duration is calculated, and the record group with a duration greater than the effective intake duration threshold is selected and the corresponding nutrition ingredient label is retained to generate effective nutrition duration data; S213: According to the duration of each type of nutrient in the effective nutrition duration data, the occurrence frequency records of each type of nutrient are counted, and an attribute set indexed by the nutrient is constructed to establish a continuous nutrition intake behavior feature set.

6. The nutritional health assessment method based on big data according to claim 1, characterized in that: The specific steps for obtaining the analysis results of the relationship between nutritional intake and health fluctuations are: S311: Acquire individual health sign data, extract measurement timestamps and corresponding values, combine the corresponding continuous intake time periods of each nutrient component in the continuous nutrition intake behavior feature set, match the health sign data covering the corresponding time period by component, and generate a sequence of corresponding health sign values within the time period; S312: Based on the continuous records of each sign in the sequence of values corresponding to the health signs within the time period, the differences between the values of adjacent records are calculated in sequence, and the differences are compared with the fluctuation thresholds respectively. All records that meet any threshold condition are marked to generate a health sign fluctuation event label set; S313: Based on the event tags identified in the health sign fluctuation event tag set, combine and merge them according to nutritional components and physical sign types, extract the number of fluctuation events in each combination and record the nutritional component index corresponding to the distribution, and establish a numerical result of the relationship between nutritional intake and health fluctuations.

7. The nutritional health assessment method based on big data according to claim 1, characterized in that: The specific steps for obtaining the list of nutrients associated with health status fluctuations are: S411: Based on the numerical results of the relationship between nutrient intake and health fluctuations, extract the time of fluctuation events and the continuous intake intervals of the nutrient components, compare the time points of each fluctuation event in chronological order to see if they fall into the time intervals of different nutrient components, determine whether the time points of fluctuation events corresponding to multiple nutrients overlap, and generate a record of health fluctuation time overlaps; S412: Based on the event records marked as time overlap in the health fluctuation time overlap records, group the occurrences of the nutrients by name and count them, retain the nutrient records with a frequency of occurrence that meets the threshold of significant associated events, and extract the corresponding intake duration to generate repeated verification stable intake data; S413: Based on the repetition times and duration values of each nutrient in the repeated verification stable intake data, a data list format is constructed with the nutrient name as the main index and the cumulative intake time as the auxiliary attribute, and various types of ingredients and their intake time are summarized and listed to establish a list of nutrients associated with health status fluctuations.

8. The nutritional health assessment method based on big data according to claim 1, characterized in that: The specific steps for obtaining the comprehensive nutrition and health assessment report are as follows: S511: Based on the index of each ingredient in the health status fluctuation-associated nutrient list, combined with the corresponding intake time segment and continuous intake time length data, a daily intake time distribution table is constructed by ingredient name, and the coverage quantity of each ingredient in the five time periods is counted to establish a nutrient time distribution record; S512: Based on the coverage amount of all ingredients in the nutrient component time distribution record in each time period, combined with the daily intake duration and intake frequency, calculate the balance coefficient of each ingredient intake behavior, and summarize and establish a nutrient intake balance index; S513: Based on the nutritional intake balance index and the number and duration of physical sign fluctuations recorded in the nutritional component list associated with the health status fluctuation, a three-dimensional data matrix indexed by components is constructed to generate a comprehensive nutritional health assessment report.

9. The nutritional health assessment method based on big data according to claim 8, characterized in that: The equalization coefficient adopts the formula: ； Calculate, where Indicates the The balance coefficient of intake behavior of various nutrients, Indicates the The nutrients in The number of coverages within a time period, Indicates the The frequency of intake of the nutrients during the statistical period, Indicates the The total duration of intake of the nutrients during the statistical period, represents the average duration of intake of all nutrients, It represents the average frequency of intake of all nutrients.

10. A nutrition and health assessment system based on big data, characterized in that: The system is used to implement the big data-based nutrition and health assessment method according to any one of claims 1 to 9, and the system comprises: The dietary segmentation module obtains individual dietary intake time, sets the dietary period division standard, compares the time to determine the segment label, calculates the intake frequency density per unit time period in each segment, and generates the food intake period coverage; The component sequence extraction module extracts the nutritional components and timestamps according to the food intake period coverage, determines the relationship between adjacent intervals and the nutritional metabolism time window threshold, summarizes the continuous intake information, and generates a continuous nutritional intake behavior feature set; The physical sign fluctuation matching module extracts the continuous time period of the components based on the continuous nutritional intake behavior feature set, obtains the health physical sign data records within the covered period, determines whether the difference between adjacent records exceeds the fluctuation threshold, determines the corresponding relationship between intake and fluctuation events, and generates an analysis result of the relationship between nutritional intake and health fluctuation; The abnormal component screening module counts the overlapping components based on the analysis results of the relationship between nutritional intake and health fluctuations, screens the items that meet the threshold of significant correlation event repeat verification criteria, and generates a list of nutritional components associated with health status fluctuations; The health assessment matrix module is based on the list of nutrients associated with health status fluctuations, extracts time segments, intake behaviors and fluctuation records, constructs a three-dimensional matrix, and generates a comprehensive nutrition and health assessment report.

Images