In modern automated systems (Interactive Voice Response, help chatbots, routing systems, etc.) it is very often important to be able to predict what is the most likely next step for the current user. One way of addressing this issue is using sequence algorithms such as Markov Chains.
After a quick introduction to the concept of Markov chains and Markov processes, we will explore the basics and the implementation of a simple High Order Markov chain to predict what the most likely next state in a sequence, based on previous states. We will be using anonymized real-life data of an automated system and we will try to come up with a model that can give us the most probable next state using Markov chains of different orders.
Things we will see in detail: - Mathematics and rationale behind Markov Chains; - Basic implementation of First Order Markov Chains; - Implementation of High Order Markov Chains; - Real life application of the developed model.
An undergraduate level of understanding of Linear Algebra and basic Python skills will be useful to follow the talk.
in __on sabato 21 aprile at 15:30 **See schedule**