Prediction details
Each row in the following table shows the accuracy of predictions by an algorithm for individual cells in the data set. "Overall best CC" indicates the performance of the current best algorithm for that cell.
3. V1 Natural Image Sequences - List cells - All predictions - My predictions - Download data set - Submit prediction!| Algorithm: | MID1 v4.0 |
| User: | jakem |
| Notes: | Used Maximally Informative Dimensions to simplify the stimulus using most informative filter, and predicted using line of best fit. Some used 2d MID, some only 1d, cells used varying methods of fitting to the curve based on how well it did in preliminary tests.
Ran preliminary tests by splitting known data into 7/8 and 1/8, running 7/8 through MID, and attempting to predict the last 1/8, then comparing. Hoping to get better results from this submission because it used all the stimulus data in MID. Ran into issues first submission with NaN values. Replaced with avg firing rate. Also ran into alignment issues. Which now should be solved.... maybe... Latest version fixes error where prediction algorithm ignored response values greater than 1 MID1 selectively ignores certain response values based on what values gave good results in my own 7/8 1/8 testing. The hypothesis is that values ignored that gave better values on the test data will do the same for the validation data, but we can't be sure until I submit this data. |
| Eval'd on: | 2010-08-05 15:38:16 (Prediction evaluated successfully) |
| Cell ID | My CC | Overall Best CC | |
| e0008 | 0.189 | 0.717 | |
| e0012 | 0.101 | 0.352 | |
| r0212b | 0.207 | 0.223 | |
| r0221a | 0.175 | 0.463 | |
| r0225c | 0.114 | 0.256 | |
| r0260 | 0.043 | 0.605 | |
| r0279 | 0.013 | 0.614 | |
| r0284 | 0.074 | 0.672 | |
| r0301 | 0.436 | 0.736 | |
| r0305 | 0.460 | 0.695 |
Not logged in. Log in - Create an account