RESUMO
Current food recommender systems tend to prioritize either the user's dietary preferences or the healthiness of the food, without considering the importance of personalized health requirements. To address this issue, we propose a novel approach to healthy food recommendations that takes into account the user's personalized health requirements, in addition to their dietary preferences. Our work comprises three perspectives. Firstly, we propose a collaborative recipe knowledge graph (CRKG) with millions of triplets, containing user-recipe interactions, recipe-ingredient associations, and other food-related information. Secondly, we define a score-based method for evaluating the healthiness match between recipes and user preferences. Based on these two prior perspectives, we develop a novel health-aware food recommendation model (FKGM) using knowledge graph embedding and multi-task learning. FKGM employs a knowledge-aware attention graph convolutional neural network to capture the semantic associations between users and recipes on the collaborative knowledge graph and learns the user's requirements in both preference and health by fusing the losses of these two learning tasks. We conducted experiments to demonstrate that FKGM outperformed four competing baseline models in integrating users' dietary preferences and personalized health requirements in food recommendations and performed best on the health task.
RESUMO
Rural depopulation is the most significant geographical phenomenon in rural areas during the process of urbanization. Although many studies have investigated the driving force of rural depopulation based on rural-urban migration at the macro level, the local factors, and their impact on rural depopulation from the rural areas have been not fully revealed. This paper selected the northern mountains of China's Hebei province as a study area to explore the pattern and local push factors of rural depopulation at the rural-township levels based on GeoDetector. The main findings are summarized as follows. (1) Rural depopulation varies substantially, demonstrates spatial correlation, and is distributed in clusters. From a dynamic perspective, compare that in years 2000-2010, the population growth areas during 2010-2017 have been significantly expanded, while the sharp depopulation areas and severe depopulation areas experienced shrinkage in our study area. (2) The pattern of rural depopulation is in accordance with terrain. Rural depopulation tends to be stronger in plateaus and mountains, while relatively milder in intermontane basins, hills, and piedmont plains. (3) The agricultural suitability of natural environmental and rural economic opportunities together with climate changes were the most important driving forces of rural depopulation at local levels. Location, sparse population, and inadequate public services also contributed to rural depopulation. However, the dominant driving factors are different in the different periods. Rural depopulation was mainly driven by arable land per capita and natural environmental variables in the years 2000-2010, while the population density, location, and off-farm economic opportunities played a decisive role in the years 2010-2017. (4) Rural depopulation is a complex, multi-dimensional process driven by a combination of multiple factors including different environmental factors, economic opportunities, and location. This paper reveals the push factors of rural depopulation in underdeveloped mountainous areas by a quantitative empirical approach, inspiring increased attention to the impacts of local factors and spatial correlations on rural depopulation, and has many implications for the policy design of China's rural revitalization.
