The area of Artificial Immune Systems (AIS) that bridges immunology, computer sci- ence, engineering and mathematics, has gained great interests in the last decade. One of the well known AIS, the Dendritic Cell Algorithm (DCA), has shown promising per- formance on the anomaly detection and attribution problem. A number of interesting properties of the algorithm could be the reason for successful applications in the past. However, there are several issues with the algorithm that limit its applicability and accessibility. As a result, this thesis aims to address and resolve these issues, by in- vestigating algorithmic properties of the DCA from both the theoretical and empirical perspectives. This leads to the research questions that are related to formalisation and runtime analysis, online analysis and automated pre-processing. In order to answer these questions, a formalisation of the DCA is first provided and runtime analyses of the algorithm are subsequently performed to show its low runtime complexity; an online analysis component is then proposed and tested to improve the algorithm's online de- tection capability; an automated pre-processing module is also developed and validated to enable the algorithm to automatically adapt to the underlying characteristics of a given problem domain. In summary, this work extends the original DCA to be more accessible to future users and more applicable to the problem of interest. The findings of this work provide novel contributions to the development and analysis of the DCA, as well as useful implications to the AIS community.