CISxx5-6 Intelligent Systems and Data Mining Assignment 1 “Designing an Intelligent System using Machine Learning” Students will explore Intelligent System (IS) technologies that are used for solving real-world problems. Students are expected to chose a domain problem related to their interests. For example, students with interests in Business Analytics can develop an IS for time series prediction and trading. The recognition methods discussed on the lectures can inspire other students to develop a face control system. A selected domain problem must be presented by a dataset that can be found in the Machine Learning Repository http://archive.ics.uci.edu/ml/ or in other open data resources http://en.wikipedia.org/wiki/Open_data. For face recognition problems students can use the Yale face data or other face datasets http://web.mit.edu/emeyers/www/face_databases.html. Students are expected to develop a solution capable of learning from the domain data and solving the problem most accurately. An example was given for a face recognition system implemented in Matlab. It will be helpful to consider one or more Working Hypotheses to be tested on the domain data. The examples of such hypotheses could be found in the lectures. Students can use Machine Learning techniques such as artificial neural networks, and decision trees. For implementing the IS, students can be inspired by the Matlab scripts developed for face recognition as discussed in the lectures, and use any language such as OpenCV or Python. The designed system should provide a prediction accuracy comparable with that provided by existing systems. Assignment Weight According to the unit unit handbook, the weight of this assignment is 50%. Submission Reports must be submitted via BREO in a word/pdf format. Report Structure & Marking Scheme Points will be given according to the criteria described in the Project Handbook No Content Points, max 1 Description of a domain problem 10 3 Description of Working Hypothesis 20 4 Experiments on the domain data (with evidences of experiments) 40 5 Conclusions 20 6 References 10 Total 100