Researchers generally have a lot of knowledge about the possibilities and might even be curious about some things themselves. Your project is expected to use some sort of novel approach.
With that said, I'd suggest that you start by reading up on existing decision tree techniques, learning why they work and what their flaws are, and try to find ways to overcome the flaws.
Then, once you have your improvement, it should be relatively easy to find a dataset to apply it to.
I have seen many people asking for help in data mining forums and on other websites about how to choose a good thesis topic in data mining.
I think this has advantages because these papers outline details regarding data as well -- perhaps you can use the same.
Present some papers and your idea to your prospective supervisor and he/she will make some suggestions. First, talk to your thesis advisor before committing to a project. Secondly, just analyzing a new dataset using standard techniques doesn't make for a good masters thesis.Since it is always easier to research a topic that the student cares about, students should pick the topic that they feel the most comfortable with.For a unique dissertation, students could find specific data sets that they want to analyze.If students want to write a dissertation on improving current techniques, they will need to decide which technique they want to focus on.Students could work in outlier detection, neural networks, algorithms, clustering, association rule mining or another field.At the very least, students will want to get their academic adviser's approval before they begin working on their chosen subject.Once students have obtained this approval, they can begin working on their data mining dissertation.The University of Minnesota is home to a wide range of data mining activities of interest to Data Science students.The departments associated with the MS Degree in Data Science, namely the Department of Computer Science and Engineering, Department of Electrical and Computer Engineering, School of Statistics, and the School of Public Health, together house a uniquely large variety of faculty and research in data mining and data management.Research projects involving the collection, management, analysis, and visualization of big data have ample opportunities for participation by students in the Data Science program for their capstone project.With the rise of technology, topics like data mining have been growing increasingly popular in dissertations.