search

MetOffice / XBTs_classification

Project for the classification of eXpendable Bathy Thermographs
BSD 3-Clause "New" or "Revised" License
4 stars 2 forks source link

Implement iMeta algorithm for comparison #11

Closed stevehadd closed 4 years ago

stevehadd commented 4 years ago

We want to be able to compare the results of ML based classifications with previous solutions to filling in missing metadata labels, specifically the iMeta algorithm described in this paper by Pallmer et. al. https://journals.ametsoc.org/doi/full/10.1175/JTECH-D-17-0129.1

it should be straightfoward to code up the decision tree and include that as a classification result. It can implemented as a function which is fed into the apply option of the XBT datset dataframe object as described here: https://www.geeksforgeeks.org/apply-function-to-every-row-in-a-pandas-dataframe/

stevehadd commented 4 years ago

I'm happy that the function for iMeta classification now works correctly. There seems to be a discrepancy in the number of labelled (model and manufacturer) profiles in the dataset I am working with, compared to the dataset in the paper, so the plots of frequency and accuracy don't look right. This needs to investigated further, so I have created a new issue (#20), and am closing this one.