Abstract
Data streams are transmitted at high speeds with huge volume and may contain critical information need processing in real-time. Hence, to reduce computational cost and time, the system may apply a feature selection algorithm. However, this is not a trivial task due to the concept drift. In this work, we show that two feature selection algorithms, Information Gain and Online Feature Selection, present lower performance when compared to classification tasks without feature selection. Both algorithms presented more relevant results in one distinct scenario each, showing final accuracies up to 14% higher. The experiments using both real and artificial datasets present a potential for using these methods due to their better adaptability in some concept drift situations.
Highlights
According to [Gradvohl 2016], Complex Event Processing (CEP) systems are stream processing distributed systems, which take one or more linearly ordered sequence of events as an input and produces another ordered sequence of events as output
Results showed that Information Gain (IG) presents the worst response time and memory consumption in all evaluated scenarios
Considering prequential accuracy, IG shows the better adaptability for gradual and sudden drifts in high dimensionality data streams, and in recurrent drifts in both high and low dimensionality. This fact demonstrates that there is a potential for using this method in concept drift environments, especially in high dimensional data streams, if we reduced the response time and memory consumption
Summary
According to [Gradvohl 2016], Complex Event Processing (CEP) systems are stream processing distributed systems, which take one or more linearly ordered sequence of events (known as Data Streams) as an input and produces another ordered sequence of events as output Another common concept is the Stream Paradigm (SP), which [Andrade et al 2011] state as a distributed computational model, supporting the continuous, heterogeneous, real-time collection and analysis of data streams. Besides its volume and velocity, one of the main challenges in learning from data streams is the dynamically changing, or non-stationary environment, which means data distribution can change over time This phenomenon, known as Concept Drift [Jankowski et al 2016], imposes difficulties for building useful solutions since every model, application or algorithm needs to adapt to different changes in data distribution
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
