Analysis of quasi-periodic space-time non-separable processes to support decision-making in medical monitoring systems

Abstract

In many decisionsupport systemsthere are processedchaotic spatial-time processes which are non-separable and quasi-periodic. Some examples of such systemsareepidemic spreading, population development, fire spreading, radio wave signals, image processing, information encryption, radio vision, etc. Processes in these systems have periodic character, e.g. seasonal fluctuations(epidemic spreading, population development), harmonic fluctuations (pattern recognition, image processing),etc. In simulation block the existing systems use separable process models which are presented as multiplication of spatialand temporal parts and are linearized. This significantly reduces the quality of spatial-time non-separable processes. The quality model building of chaotic spa-tial-time non-separable processwhich is processed by decisionsupport systemis necessary for getting of learning set. Itis really complicated especially if the random process is formed. The implementation ensemble of chaotic spatial-time non-separable process requires high costs what causes reduction of the system efficiency. Moreover, in many cases the implementation ensemble of spatial-time processes is impossible to get. In this workthemathematical model of a quasi-periodic spatial-time non-separable process has been developed. Based on it the formation method of this process has been developed and investigated. The epidemic spreading pro-cessed was presented as an example

Recommender systems: models, challenges and opportunities

“…Research has been intensively focused on improving the effectiveness of collaborative filtering in recent times. This includes the improvement of techniques for assessing similarity, the use of customer input, the increase of the quantity of data on user preferences, and other factors [29,30], [31].…”