Making Predictions with a Classifier

In order to make predictions, we need to choose a classification model. 0:00 And we're going to use the decision tree classifier that we looked at earlier. 0:05 Back in your Python file, let's add to the program on the next few lines. 0:10 First, we'll import scikit-learn's module, 0:15 which includes all of the decision tree models. 0:20 So we'll type from sklearn import tree. 0:27 Next, we'll create a decision tree classifier and 0:34 assign it to a variable so we can work with it. 0:38 We'll type classifier as the name of my variable, and we'll set that 0:42 to tree.DecisionTreeClassifier and 0:48 that's a function, so add parenthesis at the end. 0:56 Now, we need to actually build the DecisionTree through which 1:00 each new example will flow. 1:05 This decision tree can be built by feeding it both the training example and 1:08 the target labels using the fit function, like this. 1:13 So, again, we'll use our classifier variable and 1:19 we'll type classifier.fit 1:27 And inside of this function, we'll pass the Iris.data, 1:34 and then the iris.target or the labels. 1:41 So just to review, we've created a decision tree model. 1:46 And now we're actually building the decision tree using its fit function 1:52 which takes a set of examples and the target labels. 1:56 For more on the fit function, check out the notes associated with this video. 2:00 At this point the data is loaded and we've built a decision tree based on that data. 2:05 Now we can feed a new example into the top of that decision tree, and 2:10 it will flow through each branching decision like a flow chart 2:15 until it finally reaches its target label. 2:19 Now finally comes the part we've been working toward, making predictions. 2:24 We can do this using the predict function on the decision tree classifier, 2:28 like this. 2:34 So first, this is something I'm going to want to print out. 2:36 And inside of the print function, I'll type classifier.predict, 2:39 and we'll open and close some parentheses. 2:47 I am wrapping this in a print function just so 2:52 that we can see the outcome of the code when we run it. 2:54 Inside of the predict parentheses, 2:58 create two sets of nested square brackets, like this. 3:02 So there's the first pair. 3:08 And then inside of those square brackets, we'll make another pair of opening and 3:10 closing square brackets. 3:15 The outermost set of square brackets is an array of our examples. 3:18 The innermost set of square brackets 3:24 is where we'll put the values of features for a single example. 3:27 So in other words, we could predict multiple examples at a time, but 3:33 we're just sticking with one for now. 3:38 So let's start out by testing the model to make sure it's working correctly. 3:40 We can do that by just putting in an example from the data set. 3:47 From the Wikipedia page, 3:52 I'll type in the first example from the data set which happens to be a setosa. 3:53 So inside of the innermost square brackets, 4:01 I'll type 5.1, 3.5, 1.4, and 0.2. 4:06 And if we go back to the Wikipedia page to look at that, 4:12 again, you can see that this first example should be a setosa. 4:18 And now let's save it and then back in the terminal, we'll run the code. 4:27 So I'll just hit the up arrow to get the previous command and hit Enter. 4:35 And the output should be the names of the flowers because we still have 4:42 the Iris.target_names printing first. 4:46 And then next, we have this index (0), which is exactly what we want. 4:50 Remember, arrays start counting indices at 0. 4:57 And, in this case, the index is referring to these labels. 5:01 A setosa is 0, Versicolor is 1, and virginica is 2. 5:06 So zero is indeed a setosa, and so 5:13 we know that the model is predicting this correctly. 5:16 So now let's mess with this data a little bit. 5:21 In the data set, most of the setosas have a pedal width of about 0.2. 5:24 With examples ranging from 0.1 up to 0.6. 5:31 However, versicolors range from 1.0 to 1.8. 5:36 And then virginicas have petal widths from 1.4 to 2.5. 5:41 So in our example, let's change this last 5:46 feature to something like 1.5 instead. 5:51 That would be well above normal for a setosa, but 5:57 match the upper end of versicolors and just barely make the cut of virginicas. 6:01 Now save the code, and let's run it again. 6:08 You should see either a 0 or a 1. 6:15 If you hit the up arrow and hit enter, again and 6:19 again to execute the same code, you should see both numbers appearing. 6:23 That's because the decision tree classifier randomly chooses a feature 6:31 that it thinks will make the best comparison. 6:35 However, because we're always working with probabilistic behavior in machine 6:38 learning, we wont necessarily get the same result on every run. 6:43 This can indicate a low level of confidence which make sense. 6:48 The other values in our new example 6:53 don't line up with any other examples in the data set. 6:57 And we've pushed the pedal width enough 7:00 that the classifier can't really draw any confident conclusion. 7:03 It's somewhere between a setosa and a versicolor. 7:07 That's it for coding. 7:13 In our next video, we'll review some of the big ideas we've learned. 7:14