Abstract:
Most users of big-data analytics don’t know how dimensionality reduction works. Understanding the quality of a low-dimensional embedding helps choose the best dimensionality reduction algorithm and trust the transformed data.
Existing research mostly explores embeddings visually, so algorithms need to be more interpretable. We propose two interactive explanation methods for low-dimensional embeddings from any dimensionality reduction algorithm to close this gap.
LAPS generates interpretable explanations on the preserved locality for a single instance by localizing the neighborhood structure. The second method, GAPS, uses non-redundant local-approximations from a coarse discretization of the projection space to explain the retained global structure of a high-dimensional dataset in its embedding.
We apply the proposed methods to 16 real-world tabular, text, image, and audio datasets. Our extensive experimental evaluation shows that the proposed techniques can interpret low-dimensional embedding quality and select the best dimensionality reduction algorithm for any dataset.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
