Double Digits! Crazy that we are 10 weeks in!
This week we read about PCA and SVD some more, and at this weeks meeting we discussed the best way to fill in our data. Our data set is a lot of columns but only around 30 rows, due to this we are not sure we will need to reduce the dimensionality of the matrix.
Also our data set will have a lot of columns that will be empty because the “viewer” did not get to see those deals, so we are wondering which would be the best way to “fill” in the data: a random generator, K-nearest neighbor, Mutual nearest neighbor, artificial neural networks? Its still in the air.
Moving forward:
- Continue to read, read and read.
- Play around with data sets.
- Try to find the best technique to “fill” the empty spots.
- Decide if deminsionality reduction is necessary.
“Time moves slowly, but passes quickly” – Alice Walker
Bonus GRE word:
Ephemeral (adj). – short-lived