Search

Background: Reverse nearest neighbors query is an important means to solve many practical applications based on the concept of Influence Sets. It is widely used in various fields such as data mining, decision support, resources allocation, knowledge discovery, data flow, bioinformatics and so on. Objective: This work aims to improve time efficiency of Reverse Nearest Neighbors query of moving objects with large data scale. Methods: A new spatio - temporal index HRNN-tree is developed.Then an algorithm for reverse nearest neighbors query based on HRNN-tree is developed. Results: Our algorithm is superior to the existing method in execution time. The performance of our algorithm is excellent especially for the queries with large data scale and small values of k. Conclusion: This study devises a new spatio - temporal index HRNN-tree. Then an algorithm for reverse nearest neighbor search of moving objects is developed based on this index. This algorithm avoids that the query performance deteriorates rapidly as the data space grows and has a better performance fort the large data space. This work will be helpful to enrich and improve the abilities of intelligent analysis, mobile computing and quantitative query based on distance for spatio - temporal database.

Background: At present, the gain of most ILC algorithms is fixed, and the convergence speed of the system depends on the learning law, which will lead to the complexity of the structure of the learning law, and variable gain can accelerate the convergence speed without changing the structure of the learning law as variable gains are introduced into ILC. Objective: In this paper, the D-type learning law is used. Firstly, the variable gain iterative learning controller is designed. Secondly, the convergence of the learning law is analyzed. Methods: Finally, in order to illustrate the effectiveness of this method, the simulation is carried out using MATLAB. Results and Conclusion: The simulation results show that the variable gain iterative learning control can improve the convergence speed of the iteration, and weaken the restrictions on the initial input.

College Students’ mental health is an important part of higher education, so the current research and prediction of College Students’ mental health are of great significance to better solve the problem of College Students’ mental health. Taking a local university as an example, the data from 2011 to 2019 are selected and analyzed. The normalized data processing method is used to assign weights to 11 kinds of factors that affect the health of college students. The training samples of a neural network are selected, and the structural characteristics of the neural network and the artificial neural network toolbox of MATLAB are used to establish the BP based model the mathematical model of the prediction system of College Students’ mental health based on neural network. The results show that the error between the predicted value and the measured value is only 0.88%. On this basis, this paper uses the model to predict the weight of the influencing factors of the mental health status of college students in a local university in 2020 and analyzes the causes of the prediction results, to provide the basis for the current mental health education of college students.