Python Machine Learning Cookbook - Second Edition

How to do it...

Let's take a look at how to implement ranking algorithms:

1. Create a new Python file and import the following package (the full code is in the LambdaMARTModel.py file that's already provided for you):

import pyltr

2. We will load the data contained in the Letor dataset that's already provided for you (train.txt, vali.txt, and test.txt):

with open('train.txt') as trainfile, \        open('vali.txt') as valifile, \        open('test.txt') as testfile:    TrainX, Trainy, Trainqids, _ = pyltr.data.letor.read_dataset(trainfile)    ValX, Valy, Valqids, _ = pyltr.data.letor.read_dataset(valifile)    TestX, Testy, Testqids, _ = pyltr.data.letor.read_dataset(testfile)    metric = pyltr.metrics.NDCG(k=10)

3. Let's now perform a validation of the data:

 monitor ...