Closed jovo closed 8 years ago
sections:
always everything only on half the data data which dataset definition of markers how did puncta get to us (what computer vision) definition of features (with equations) feature exploration synapsin1 vs. synapsin2 in log domain and linear same for vglut1 & 2 repeat the above 4 panel figure for each of the features kde plots of chosen transformation/feature pair marker exploration correlation matrix 2D scatter plot colored by truth 2D scatter plot colored by truth, overlay optimal voronoi diagram (fit using linear discriminant analysis, not kmeans) same as above in 3D ARI vs. dimension using optimal voronoi diagram synapse exploration kmeans (k=2) heatmap lattice plots of GABABR (fix colors) correlation matrices for each of the 2 clusters kmeans (k=2) for level 2
always everything only on half the data
i think this is a cool plan. what do you think?
That's a good looking outline.
The skeleton is up.
Moved issue to weekly-status-reports.
sections: