top of page
Thesis Title:
Graph-based Learning using Multiple Views
Work Details:
1. Represent the data set as a graph and then learn the optimal graph similarity matrix that can divide the data set into some clusters and can also perform semi-supervised classification on the given data set.
2. Use multiple views or multiple distinct feature sets of data set which give dierent partial information about a data set thus improving the graph-based learning performance.
3. Incorporate the kernel method to consider the nonlinearity present in the data set and to solve the issue of choice of kernel, multiple kernel learning is also incorporated to improve the learning performance.
4. Development of several robust graph-based learning algorithm by integrating multiple views and multiple kernel learning.
bottom of page