A Knowledge-based Recommendation System for Time Series Classification

Time series data sets reflect the state and extent of things as they change over time. Information extraction based on such data plays an important role in many fields. The time series classification is a typical supervised learning problem, which is applied in speech recognition, image processing a...

Full description

Bibliographic Details
Main Authors: Man Tianxing, Nataly Tianxing, Nikolay Mustafin
Format: Article
Language:English
Published: FRUCT 2019-04-01
Series:Proceedings of the XXth Conference of Open Innovations Association FRUCT
Subjects:
Online Access:https://fruct.org/publications/abstract24/files/Tia.pdf
Description
Summary:Time series data sets reflect the state and extent of things as they change over time. Information extraction based on such data plays an important role in many fields. The time series classification is a typical supervised learning problem, which is applied in speech recognition, image processing and so on. However, because the attributes of time series data don't make sense and the feature dimensions are particularly large, people can't treat them as general machine learning classification problems. Currently, many different time series classification problems have been proposed. But how to choose and use these methods is still a huge problem for non-computer professional researchers. This article uses the ontology technology to build a recommendation system that contains the details and features of such algorithms. When the users input the characteristics of the data and the task requirements, they can get reasonable suggestions and a description of the workflow of the algorithm. Such a system saves the user a lot of analysis and comparison time. It also makes such problems easier to understand.
ISSN:2305-7254
2343-0737