Detecting Noise in Fractal Time Series Using Machine Learning

This study concentrates on devising a method to evaluate the level of noise in fractal Brownian motion through machine learning methods. A method for classifying trajectories of fractal Brownian motion with varying levels of additive noise using a convolutional neural network has been proposed. Modeled fractal time series with additive noise were utilized as the input dataset. The noise component was generated with different dispersion values, allowing the investigation of the noise level's influence on the system and its environment. The results provide insights into the effectiveness and trustworthiness of employing these machine learning techniques for assessing noise within fractal systems.