Our society is becoming more digitalized for each day. Now, we are able to gather data from individual users with higher resolution than ever. With the increased amount of data on an individual user level, we can analyze their behavior. This is of interest in many different domains, for example service providers wanting to improve their service for their customers. If they know how their service is used, they have more insight in how they can improve. But, it also imposes additional difficulties. When we reach the individual user, the irregularities in the regular behavior makes it harder to model the normal behavior.

In this thesis, we explore data-driven techniques to model and analyze user behaviors. We aim to evaluate existing as well as develop novel technologies to identify approaches that are suitable for use on an individual user level. We use both supervised and unsupervised learning methods to model the user behavior and evaluate the approaches on real world electricity consumption data.

Firstly, we analyze household electricity consumption data and investigate the use of regression to model the household's behavior. We identify consumption trends, how data granularity affects modeling, and we show that regression is a viable approach to model user behavior. Secondly, we use clustering analysis to profile individual households in terms of their electricity consumption. We compare two dissimilarity measures, how they affect the clustering analysis, and we investigate how the produced clustering solutions differ. Thirdly, we propose a sequential clustering algorithm to model evolving user behavior. We evaluate the proposed algorithm on electricity consumption data and show how the produced model can be used to identify and trace changes in the user's behavior. The algorithm is robust to evolving behaviors and handles both dynamic and incremental aspects of streaming data.

Streaming data is becoming more prevalent in our society every day. With the increasing use of technologies such as the Internet of Things (IoT) and 5G networks, the number of possible data sources steadily increases. Therefore, there is a need to develop algorithms that can handle the massive amount of data we now generate.

Data mining is an area of research where data is mined to gain an understanding of data and its underlying structure. When we move to streaming data, and the corresponding sub-domain data stream mining, restrictions are imposed on the algorithms that can be used. Data streams are possibly endless, and their instances arrive rapidly, can often only be processed once or a few times, and often evolve as the data is generated over time.

This thesis explores data-driven techniques to model and analyze evolving data streams. We focus on slower data streams where incremental updates are not necessary, and the interest lies in analyzing its behavior over longer time periods. We aim to evaluate existing and develop novel algorithms and techniques suitable for analyzing these types of data streams. We use both supervised and unsupervised learning methods to model the user/system behaviors, and the methods and algorithms are evaluated on various datasets.

Specifically, we investigate regression and clustering algorithms to mine streaming data for user/system behavior patterns. We also design an algorithm capable of modeling user/system behavior in a single evolving data stream, which is easy to use and capitalizes on prior knowledge from the history of the stream. Furthermore, we design a clustering algorithm that takes advantage of multiple data streams, where each stream represents a part of the entire system, to model various aspects of the user/system behavior. Finally, we review the current state-of-the-art methods for evaluating data stream clustering algorithms and identify aspects that should be considered for the future.