Dana Honeycutt has just contributedPCA vs tSNE for Breast Cancer Datato the ScienceCloud Exchange. |
The ScienceCloud Exchange is a community that brings together Pipeline Pilot developers, softwareproviders, and Pipeline Pilot end-users to share Pipeline Pilot protocols and accelerate the pace of innovation.
| Name: | PCA vs tSNE for Breast Cancer Data | |
| Summary: | Compares PCA to t-SNE for breast cancer data | |
| Description: | t-Distributed Stochastic Neighbor Embedding (t-SNE) is a dimensionality reduction technique useful for visualization of small to medium-sized data sets. Here, we compare it to the classic PCA method. Visualizations are shown using both dimensionality reduction techniques in order to determine which approach best separates benign (green) from malignant (red) cases. Results for reduction to both 3 dimensions and 2 dimensions are shown. In both cases, t-SNE gives better separation between the two classes, though the t-SNE computations are slower compared to PCA. Note: Requires Pipeline Pilot 2018 and the Analytics and Machine Learning Collection, with PP configured for R usage and the required R packages installed. #Visualization#Reporting#Statistics#R#Generic | |
