| CPC G06F 18/22 (2023.01) [G06F 18/23 (2023.01); G06F 18/24147 (2023.01); G16Y 40/20 (2020.01); G16Y 10/60 (2020.01); G16Y 20/10 (2020.01); G16Y 40/10 (2020.01)] | 11 Claims |
|
1. A method for construction and management of children's management institutions in a smart city area based on an Internet of Things (IoT), wherein the method is realized based on a system for construction and management of children's management institutions in the smart city area based on the IoT, the system comprising a user platform, a service platform, a management platform, a sensor network platform, and an object platform, the system being a part of a server or implemented by the server, the user platform including a first terminal device and being used to obtain user needs, the service platform being used to communicate between the user platform and the management platform for providing users with input and output services, the management platform being configured as a second terminal device and used to achieve connection and collaboration between the user platform, the service platform, the sensor network platform, and the object platform, the sensor network platform being used to connect the management platform and the object platform for achieving functions of sensing information sensor communication and control information sensor communication, and the object platform being used for generation of sensor information and execution of control information, the method being executed by the management platform, the method comprising:
obtaining population-related data in a historical target time period through the object platform based on the sensor network platform;
predicting a number of children population in a future target time period based on the population-related data, wherein the population-related data includes data related to children population prediction within a certain area, wherein the population-related data includes at least one of marriage data, pregnant women file data, registration data of birth population, infant data in a nursery, and infant vaccination data;
wherein the predicting a number of children population in a future target time period based on the population-related data includes:
predicting a number of infant population of a target area in a first future time period based on the population-related data of the target area in the historical target time period; and predicting a number of school-age children population of the target area in a second future time period based on the population-related data; including:
obtaining a first population feature vector through processing the population-related data of the target area by a first embedding layer, wherein the first embedding layer is a model for extracting the first population feature vector, the first embedding layer includes a long short term memory network (LSTM) model, wherein an input of the first embedding layer includes the population-related data of the target area, and an output of the first embedding layer includes the first population feature vector of the target area, and the first population feature vector is a vector corresponding to a population change feature of the target area;
obtaining at least one second population feature vector through processing population-related data of at least one sample area in the historical target time period by a second embedding layer, wherein the second embedding layer includes the LSTM model, and the second population feature vector is a vector corresponding to a population number change feature in the sample area; wherein elements of the first population feature vector or the second population feature vector include a growth number and a proportion of infants born, a growth number and a proportion of immigrant infants, a growth number and a proportion of school-age children born, and a growth number and a proportion of immigrant school-age children, wherein
the first embedding layer and the second embedding layer are obtained through a first training process based on a plurality of first training samples with first labels, wherein the first training samples include at least data related to the birth population of different sample areas in a same historical target time period, and basic development data corresponding to different sample areas in the same historical target time period, and the first labels include a similarity between a proportion of an increase in a number of school-age children born in the sample area, and a similarity between a proportion of an increase in a number of school-age children in the sample area, and the first labels are determined by calculating an Euclidean distance between the first population feature vector and the second population feature vector;
the first training process includes: inputting the plurality of first training samples with first labels into an initial first embedding layer or an initial second embedding layer, constructing a first loss function based on the first labels and output results of the initial first embedding layer or the initial second embedding layer, updating parameters of the initial first embedding layer or the initial second embedding layer based on the first loss function; and obtaining the first embedding layer or the second embedding layer until the first loss function of the initial first embedding layer or the initial second embedding layer meeting a first preset condition, wherein the first preset condition includes a convergence of the first loss function and a number of iterations reaching a first threshold;
obtaining at least one cluster center through clustering the at least one second population feature vector based on a cluster algorithm, wherein the cluster algorithm includes one of a K-means algorithm, density clustering, hierarchical clustering, Gaussian mixture clustering;
taking an average growth rate of the infants corresponding to a cluster center closest to the first population feature vector as a growth rate of the infants of the target area; and
taking an average growth rate of the school-age children corresponding to a cluster center closest to the first population feature vector as a growth rate of the school-age children of the target area;
determining a construction plan of the children's management institutions based on the number of children population, wherein the construction plan includes a number of the children's management institutions; and
feeding the construction plan back to a user through the user platform based on the service platform.
|