Analysis of Artificial Intelligence Methods and Algorithms for Processing Data as a Series of Signals

The paper deals with the issues of improving the accuracy when performing the task of classifying several signals, where the main task is to determine the class of an object based on the data of the time series of this object. Examples of such signals are ECG signals, sounds, vibrations, and others....

Full description

Saved in:
Bibliographic Details
Published in2023 XXVI International Conference on Soft Computing and Measurements (SCM) pp. 152 - 155
Main Authors Belyaev, Pavel Y., Sheinman, Elena L., Kim, Iuliia V.
Format Conference Proceeding
LanguageEnglish
Published IEEE 24.05.2023
Subjects
data analysis
Data models
machine learning
Machine learning algorithms
signal processing
Sonar
Sonar measurements
Support vector machines
Training
Vibrations
Online AccessGet full text

Cover

More Information
Summary:The paper deals with the issues of improving the accuracy when performing the task of classifying several signals, where the main task is to determine the class of an object based on the data of the time series of this object. Examples of such signals are ECG signals, sounds, vibrations, and others. To successfully solve the classification problem, it is important to select the appropriate method correctly and qualitatively prepare the data for model training, including pre-processing of data and selection of model parameters. This study includes a review of artificial intelligence methods in the field of data analysis based on several signals, including machine learning algorithms. The datasets used for the study are Sonar, Doppler, and Winnipeg. Based on the comparison of the studied methods, SVM, Random Forest, AdaBoost, KNN showed the highest accuracy. The average accuracy of the classification was 0.9.
DOI:10.1109/SCM58628.2023.10159109