| Latent Semantic Analysis (LSA) Tutorial - Part 4 - Clustering by Color |
 |
 |
 |
|
Page 6 of 8
Part 4 - Clustering by ColorWe can also turn the numbers into colors. For instance, here is a color display that corresponds to the first 3 dimensions of the Titles matrix that we showed above. It contains exactly the same information, except that blue shows negative numbers, red shows positive numbers, and numbers close to 0 are white. For example, Title 9, which is strongly positive in all 3 dimensions, is also strongly red in all 3 dimensions.
We can use these colors to cluster the titles. We ignore the first dimension for clustering because all titles are red. In the second dimension, we have the following result.
Using the third dimension, we can split each of these groups again the same way. For example, looking at the third dimension, title 6 is blue, but title 7 and title 9 are still red. Doing this for both groups, we end up with these 4 groups.
It’s interesting to compare this table with what we get when we graph the results in the next section.
|
