Non-Gaussian Feature Distribution Forecasting Based on ConvLSTM Neural Network for Robust Machine Condition Prognosis

Explore a 25-minute conference talk from the 51st Conference on Applications of Mathematics, presented by Dawid Szarek from Wrocław University of Science and Technology. Delve into the topic of Non-Gaussian feature distribution forecasting using ConvLSTM neural networks and its application to robust machine condition prognosis. Learn about the problem statement, key components, input vector optimization, and outside tasks related to this advanced forecasting technique. Compare different approaches, evaluate the results, and examine practical examples. Gain insights into the conclusions drawn from this research, which combines deep learning and statistical analysis for improved machine condition monitoring and prediction.

Introduction
Problem Statement
Components
Input Vector
Optimization
Outside task
Comparison
Evaluation
Results
Examples
Conclusions