CN-122025018-A - Particle swarm optimization-based nutrition catering recommendation system and method

CN122025018ACN 122025018 ACN122025018 ACN 122025018ACN-122025018-A

Abstract

The invention relates to the technical field of data analysis, in particular to a nutrition match recommendation system and method based on a particle swarm algorithm, wherein the method comprises the steps of collecting nutrition matches and corresponding characteristics of a user; the method comprises the steps of establishing mass preference nutrition matching meal, counting the time length of a user using a nutrition matching meal recommendation system, comparing characteristics with mass preference nutrition matching meal, determining the preference specificity degree of the user, analyzing the characteristics, adjusting the cognition coefficient of a particle swarm algorithm in combination with the preference specificity degree, constructing a characteristic intensity sequence based on the characteristics, analyzing the characteristic intensity sequence to obtain recommendation weights of the characteristic intensity, correcting fitness function weights of each characteristic according to the recommendation weights of the characteristic and the characteristic intensity, outputting the recommendation nutrition matching meal corresponding to the user by adopting the particle swarm algorithm to synthesize the adjusted cognition coefficient, the recommendation weights of the characteristic intensity and the corrected fitness function weights, visually displaying, comprehensively determining all parameters in the algorithm, and recommending the nutrition matching meal which meets requirements to the user.

Inventors

WU XIKUN
GUO XIAOZHENG
BAI MUHAI
XIAO XIONG

Assignees

北京科原健康科技有限公司

Dates

Publication Date: 20260512
Application Date: 20251231

Claims (10)

1. The nutrition catering recommendation method based on the particle swarm optimization is characterized by comprising the following steps: Acquiring a nutrition matching meal recommendation system, and acquiring nutrition matching meal and corresponding characteristics of a user; Establishing a public preference nutrition catering, counting the time length of using a nutrition catering recommendation system by a user, comparing the characteristics with the public preference nutrition catering, and determining the preference specificity degree of the user; analyzing the characteristics, and adjusting the cognitive coefficient of the particle swarm algorithm by combining the preference specificity degree; constructing a characteristic intensity sequence based on the characteristics, and analyzing the characteristic intensity sequence to obtain recommended weight of the characteristic intensity; modifying the fitness function weight of each feature according to the recommended weights of the features and the feature intensities; And outputting the recommended nutrition meal corresponding to the user by adopting the particle swarm algorithm to comprehensively adjust the cognitive coefficient, the recommended weight of the characteristic intensity and the corrected fitness function weight, and visually displaying.
2. The method for recommending nutritional ingredients based on particle swarm optimization according to claim 1, wherein the steps of acquiring a nutritional ingredient recommendation system, acquiring a nutritional ingredient of a user and corresponding characteristics, and comprising: the nutrition match recommending system formulates corresponding nutrition matches according to different user demands, quantifies each characteristic in each nutrition match and records all intensities of the characteristics correspondingly; And generating a corresponding trigger event according to the operation of selecting the nutrition match by the user, and recording the user, the nutrition match, the characteristics, the selection occurrence time and the selection occurrence information corresponding to the trigger event.
3. The method of claim 2, wherein the characteristics include, but are not limited to, nutritional ingredients, food material composition, and food material taste.
4. The method for recommending nutritional ingredients based on particle swarm optimization according to claim 2, wherein establishing a popular preferred nutritional ingredient, counting the time period of using the nutritional ingredient recommendation system by the user, comparing the characteristics with the popular preferred nutritional ingredient, and determining the preference specificity degree of the user comprises: Establishing mass preference nutrition matching meal according to the history record of the nutrition matching meal recommendation system by adopting a particle swarm algorithm; Counting the time length of using the nutrition catering recommendation system by all users, and screening the maximum time length; Comparing the characteristics with the public preference nutrition match to obtain the existing intensity difference value of the nutrition match corresponding to the current user and the public preference nutrition match in each characteristic, determining the average value of the existing intensity difference values of each characteristic, and screening the largest existing intensity difference value of the characteristics corresponding to the average value of the existing intensity difference values; and determining the preference specificity degree of the current user by integrating the duration, the duration maximum value, the existing intensity difference average value and the maximum existing intensity difference value.
5. The particle swarm algorithm-based nutritional meal recommendation method according to claim 2, wherein analyzing the characteristics, and adjusting the cognitive coefficients of the particle swarm algorithm in combination with the preference specificity comprises: analyzing the characteristics, counting the intensity proportion situation of the characteristics in the nutrition catering of the user, and evaluating the increase degree of the cognitive coefficient by combining the preference specificity degree; and determining a reference cognitive coefficient according to the mass preference nutrition catering, and adjusting the cognitive coefficient of the user at the current moment by combining the reference cognitive coefficient and the increase degree of the cognitive coefficient.
6. The particle swarm optimization-based nutritional meal recommendation method according to claim 5, wherein analyzing the characteristics, counting the intensity ratio of the characteristics in the nutritional meal of the user, and evaluating the increase degree of the cognitive coefficient in combination with the preference specificity degree comprises: Based on the characteristics, counting the intensity type proportion of each characteristic in the nutrition catering of the current user, obtaining the average value of the intensity types proportion of all the characteristics, and screening the minimum value of the intensity types proportion of the characteristics in all the nutrition catering; The intensity type of other users in the nutrition recipe recommendation system is acquired by the same method and the intensity type is screened to obtain the maximum value of the average value of the intensity types; and obtaining the increase degree of the cognitive coefficient by combining the preference specificity degree through the intensity type of the current user accounting for the average value and the minimum value and the intensity type of other users accounting for the maximum value of the average value.
7. The particle swarm optimization-based nutritional meal recommendation method according to claim 5, wherein the constructing a characteristic intensity sequence based on the characteristics, analyzing the characteristic intensity sequence to obtain a recommendation weight of the characteristic intensity, comprises: Sorting according to the selection occurrence time corresponding to the intensity of each feature in the nutrition match of the user to establish a feature intensity sequence; counting the intensity occurrence of each feature according to the nutrition catering of the user, and acquiring the intensity distribution condition of each feature based on the feature intensity sequence; Evaluating the recommended weight change degree of the intensity of each feature by combining the intensity occurrence condition, the intensity distribution condition and the increase degree of the cognitive coefficient; And determining a reference recommendation weight according to the public preference nutrition catering, and obtaining the recommendation weight of the characteristic intensity at the current moment through the reference recommendation weight and the recommendation weight change degree.
8. The nutrition recipe recommendation method based on the particle swarm optimization according to claim 7, wherein the obtaining of the intensity distribution of each feature based on the feature intensity sequence is specifically: And defining any intensity in the characteristic intensity sequence as target intensity, dividing the region containing the target intensity in the characteristic intensity sequence into a plurality of divided regions, determining the distribution density of the target intensity in each divided region according to the intensity quantity in the divided regions, and acquiring the distribution density corresponding to the nearest divided region of the divided region corresponding to the current moment.
9. The particle swarm optimization-based nutritional meal recommendation method according to claim 5, wherein the modifying the fitness function weight for each feature according to the recommended weights of the feature and the feature strength comprises: equalizing the recommended weight of the characteristic intensity, and comparing the recommended weight with the recommended weight average value to obtain the recommended difference expression of the characteristic; the difference expression and the intensity ratio situation are comprehensively recommended to obtain the change degree of the fitness function weight of the current moment characteristic; Determining a reference fitness function according to the public preference nutrition catering, and obtaining a fitness function proportion value of the current moment characteristic through the reference fitness function and the fitness function weight change degree; and dynamically adjusting the fitness function weight of the feature by using the fitness function proportional value.
10. The nutrition recipe recommendation system based on the particle swarm optimization is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are communicated with each other through the communication bus, and the processor calls logic instructions in the memory to execute the nutrition recipe recommendation method based on the particle swarm optimization according to any one of claims 1-9.

Description

Particle swarm optimization-based nutrition catering recommendation system and method Technical Field The invention relates to the technical field of data analysis, in particular to a nutrition recipe recommendation system and method based on a particle swarm algorithm. Background The nutrition meal recommendation based on the particle swarm algorithm can efficiently balance multi-objective conflicts such as nutrition, cost and taste, and the pareto optimal solution set is rapidly locked in a huge recipe space through simulating swarm intelligence, so that the behavior mode of foraging of the bird swarm is used for reference, each particle represents a possible meal distribution scheme, and the meal distribution recommendation required by a task is formulated from a large amount of food materials through the combination of global search and local fine search. The dynamic personalized core advantage of the method enables the algorithm to adaptively adjust the searching direction according to the real-time feedback of the user and the historical data, ensures that the recommendation result is scientific and accurate and fits personal preference, and realizes real 'thousands of people and thousands of sides' healthy diet planning. However, in the prior art, when the particle swarm algorithm is adopted to recommend the nutrition recipe, the recommendation weights of different characteristic intensities are only modified according to the selection proportion of the historical recipe characteristics of the user, the mutation of the user preference is not considered, the recommended recipe does not accord with the preference bias of the recent user, and the user experience is poor, for example, when the user prefers the light diet for a long time and needs to turn to the high-protein diet due to seasonal variation or physical condition adjustment, if the recommendation is still carried out according to the historical recipe characteristics, the recipe mainly comprising vegetables and low-fat food materials is still continuously pushed, the user requirement cannot be met, and the recommendation of the current nutrition recipe is disjointed with the current real requirement of the user. Disclosure of Invention In order to solve the technical problems that the nutrition matching meal recommendation in the prior art is possibly not in line with the preference bias of a recent user and the preference grasp of the user is not in time, the invention aims to provide a nutrition matching meal recommendation method based on a particle swarm algorithm, and the adopted technical scheme is as follows: Acquiring a nutrition matching meal recommendation system, and acquiring nutrition matching meal and corresponding characteristics of a user; Establishing a public preference nutrition catering, counting the time length of using a nutrition catering recommendation system by a user, comparing the characteristics with the public preference nutrition catering, and determining the preference specificity degree of the user; analyzing the characteristics, and adjusting the cognitive coefficient of the particle swarm algorithm by combining the preference specificity degree; constructing a characteristic intensity sequence based on the characteristics, and analyzing the characteristic intensity sequence to obtain recommended weight of the characteristic intensity; modifying the fitness function weight of each feature according to the recommended weights of the features and the feature intensities; And outputting the recommended nutrition meal corresponding to the user by adopting the particle swarm algorithm to comprehensively adjust the cognitive coefficient, the recommended weight of the characteristic intensity and the corrected fitness function weight, and visually displaying. Preferably, the nutrition recipe recommendation system is configured to acquire a nutrition recipe and corresponding features of a user, including: the nutrition match recommending system formulates corresponding nutrition matches according to different user demands, quantifies each characteristic in each nutrition match and records all intensities of the characteristics correspondingly; And generating a corresponding trigger event according to the operation of selecting the nutrition match by the user, and recording the user, the nutrition match, the characteristics, the selection occurrence time and the selection occurrence information corresponding to the trigger event. Preferably, the features include, but are not limited to, nutritional ingredients, food material composition, and food material taste. Preferably, establishing a public preference nutrition catering, counting the time length of using a nutrition catering recommendation system by a user, comparing characteristics with the public preference nutrition catering, and determining the preference specificity degree of the user, wherein the method comprises the following steps: Establishing mass prefe