Order from us for quality, customized work in due time of your choice.
The topic of the thesis is all about coming up with a new method of anomaly detection within time series as well as an in depth peer review of the recent methods and applications that have been used over the last 2-3 years in this field of anomaly detection such as matrix profiling and machine learning and any other relevant techniques,
these 25 pages will be the main foundation of the thesis that should cover the following:
ABSTACT / INTRODUCTION : the motivation behind investigating why time series and anomaly detection are important in various industries (Hard sell on the importance of anomaly detection), the problems and hurdles that are present in anomaly detection thats causing issues, The objective of this thesis which is to come up with a new method of anomaly detection using recent findings and advancements ( can be left vague for the time being and as general as possible), General descriiptions of what time series are as well as the different anomalies that will be delved into in the background information.
BACKGROUND INFORMATION: all the important definitions and terminologies used to introduce the concepts used within the thesis: such as Time series, Anomaly (Outlier) and their different their types Point Anomalies, Contextual Anomalies, Collective Anomalies. Time Series, Anomaly Detection, Seasonality, Trend, Noise, Feature Extraction, Unsupervised /Supervied Learning, False Positive /Negative, Precision, Recall, F1 Score, Autoencoders, Neural Networks, and any recent terminology useful to the study.
RELATED WORK – relevant JOURNAL FORMAT studies over the last 3 years (minimum 10) that cover topics such as Anomaly detection in general, Anomaly detection in Time series, Peer Review of related works that compares the pros and cons of each method in a table at the end of the section
it is encouraged to reference Eamonn Keogh with his recent relevant findings found here:
https://www.cs.ucr.edu/%7Eeamonn/selected_publications.htm
and talk about his findings on his paper on how time series are flawed (uploaded with various other important works)
online references for papers can/should include Springer, IEEE Explorer, ACM, Elsevier specifically works over the last 2-3 years as well (please include references as well as a copy of whatever refences were used in the work done)
Order from us for quality, customized work in due time of your choice.