The Virtual Machine Learning Laboratory with Visualization of Algorithms Execution Process

Abstract

This paper describes the development of a virtual laboratory for teaching various machine learning algorithms with a visualization of the execution process. The novelty of the results lies in the fact that, in contrast to existing virtual machine learning laboratories, the presented one provides the visualization system for machine learning algorithms execution process which instantly shows changes by the parameters and changes in the software implementation code. Also, the open-source structure of an algorithm provides an ability for third-party interested developers to add their own lessons that have passed the validation. Visualization of the machine learning algorithms execution process is demonstrated by the example of solving the task of finding the shortest path. The student can independently build a map of the area in 2D and 3D, dynamically change it and trying different algorithms to find the shortest path. This allows for a comparative analysis of various machine learning algorithms when it comes to spatial orientation. The developed laboratory currently has no analogues among widely available software tools for training specialists in the field of data science and machine learning.