1 00:00:18,500 --> 00:00:23,340 Hi! I went to see that movie The Great Gatsby last night. I thought that was really good. 2 00:00:23,340 --> 00:00:30,340 I hope you don't mind if I finish off my martini. Anyway, one of the constantly recurring themes 3 00:00:32,710 --> 00:00:39,280 in this course is the necessity to get close to your data, look at it in every possible 4 00:00:39,280 --> 00:00:46,280 way. In this last lesson of the first class, we're going to look at visualizing your data. 5 00:00:46,410 --> 00:00:50,449 This is what we're going to do. We're going to use the Visualize panel. I'm going to open 6 00:00:50,449 --> 00:00:57,449 the iris dataset. You came across the iris dataset in one of the activities, I think. 7 00:00:57,739 --> 00:01:04,739 I'm using it because it has numeric attributes, four numeric attributes: sepallength,sepalwidth, 8 00:01:06,210 --> 00:01:13,210 petallength, petalwidth. The class are the three kinds of iris flower: Iris-setosa, Iris-versicolor, 9 00:01:15,090 --> 00:01:21,390 and Iris-virginica. Let's go to the Visualize panel and visualize 10 00:01:21,390 --> 00:01:28,390 this data. There is a matrix of two dimensional plots, a five-by-five matrix of plots. 11 00:01:31,470 --> 00:01:40,600 If I can select one of these plots, I'm going to be looking at a plot of sepalwidth on the x-axis and petalwidth on the y-axis. 12 00:01:42,150 --> 00:01:49,150 That's a plot of the data. The colors correspond to the three classes. I can actually change 13 00:01:49,420 --> 00:01:52,920 the colors. If I don't like those, I could select another color, but I'm going to leave 14 00:01:52,920 --> 00:01:59,040 them the way they are. I can look at individual data points by clicking on them. This is talking 15 00:01:59,040 --> 00:02:06,040 about instance number 86 with a sepallength of 6, sepalwidth of 3.4, and so on. 16 00:02:07,450 --> 00:02:14,450 That's a versicolor, which is why this spot is colored red. We can look individual instances. 17 00:02:15,790 --> 00:02:20,620 We can change the x- and y-axis by changing on the menus here. Better still, if we click 18 00:02:20,620 --> 00:02:25,840 on this little set of bars here, these represent the attributes. I'm going to click on this 19 00:02:25,840 --> 00:02:32,840 and the x-axis will change to sepallength. Here the x-axis is sepalwidth. Here the x-axis 20 00:02:33,610 --> 00:02:40,610 is petallength, and so on. If I right click, then it will change the y-axis to sepallength. 21 00:02:41,590 --> 00:02:55,890 So, I can quickly browse around these different plots. There is a Jitter slider. 22 00:02:56,700 --> 00:03:01,390 Sometimes, points sit right on top of each other, and jitter just adds a little bit of randomness 23 00:03:01,390 --> 00:03:06,900 to the x- and the y-axis. With a little bit of jitter on here, the darker spots represent 24 00:03:06,900 --> 00:03:14,320 multiple instances. If I click on one of those, I can see that that point represents three 25 00:03:14,320 --> 00:03:20,020 separate instances, all of class iris-setosa, and they all have the same value of petallength 26 00:03:20,020 --> 00:03:21,990 and sepalwidth. 27 00:03:21,990 --> 00:03:29,930 Both of which are being plotted on this graph. The sepalwidth and petallength are 3.0 and 1.4 28 00:03:29,930 --> 00:03:34,210 for each of the three instances. 29 00:03:36,220 --> 00:03:43,220 If I click another one here. This one here are two with very similar [sepalwidths] and petallengths, 30 00:03:47,560 --> 00:03:49,590 both of the class versicolor. 31 00:03:49,590 --> 00:03:54,190 The jitter slider helps you distinguish between points that are in fact very close together. 32 00:03:54,190 --> 00:04:01,190 Another thing we can do is select bits of this dataset. I'm going to choose select rectangle 33 00:04:01,690 --> 00:04:08,690 here. If I draw a rectangle now, I can select these points. If I were to submit this rectangle, 34 00:04:09,450 --> 00:04:14,110 then all other points would be excluded and just these points would appear on the graph, 35 00:04:14,110 --> 00:04:21,110 with the access re-scaled appropriately. Here we go. I've submitted that rectangle, and 36 00:04:21,260 --> 00:04:26,450 you can see that there's just the red points and green points there. I could save that 37 00:04:26,450 --> 00:04:33,050 if I wanted as a different dataset, or I could reset it and maybe try another kind of selection 38 00:04:33,050 --> 00:04:37,550 like this, where I'm going to have some blue points, some red and some green points and 39 00:04:37,550 --> 00:04:43,360 see what that looks like. This might be a way of cleaning up outliers in your data, 40 00:04:43,360 --> 00:04:50,360 by selecting rectangles and saving the new dataset. 41 00:04:50,480 --> 00:04:57,480 That's visualizing the dataset itself. What about visualizing the result of a classifier. 42 00:04:58,820 --> 00:05:05,820 Let's get rid of this visualize panel and back to the Preprocess panel. I'm going to 43 00:05:07,010 --> 00:05:14,010 use a classifier. I'm going to use, guess what, J48. Let's find it under trees. I'm 44 00:05:14,430 --> 00:05:21,430 going to run it. Then if I right click on this entry here in the log area, I can view 45 00:05:25,920 --> 00:05:32,920 classifier errors. Here we've got the class plotted against the predicted class. The square 46 00:05:33,770 --> 00:05:39,300 boxes represent errors. If I click on one of these, I can, of course, change the different 47 00:05:39,300 --> 00:05:45,610 axes if I want. I can change the x-axis and the y-axis, but I'm going to go back to class 48 00:05:45,610 --> 00:05:55,710 and predictedclass. If I click on one of these boxes, I can see where the errors are. 49 00:05:57,210 --> 00:06:04,210 There are two instances where the predicted class is versicolor and the actual class is virginica. 50 00:06:04,820 --> 00:06:10,550 We can see these in the confusion matrix. The actual class is virginica, and the predicted 51 00:06:10,550 --> 00:06:17,550 class is versicolor, that's 'b'. This 2 entry in the confusion matrix is represented by these 52 00:06:17,550 --> 00:06:28,680 2 instances here. If I look at another point, say this one. Here I've got one instance which 53 00:06:28,680 --> 00:06:41,290 is in fact a setosa predicted to be a versicolor. I can look at this plot and find out where 54 00:06:41,290 --> 00:06:48,290 the misclassifications are actually occurring, the errors in the confusion matrix. 55 00:06:51,150 --> 00:06:56,710 Get down and dirty with your data and visualize it. You can do all sorts of things. You can 56 00:06:56,710 --> 00:07:01,150 clean it up, detect outliers. You can look at the classification errors. For example, 57 00:07:01,150 --> 00:07:06,390 there's a filter that allows you to add the classifications as a new attribute. 58 00:07:06,390 --> 00:07:12,150 Let's just go and have a look at that. I'm going to go and find a filter. We're going to add 59 00:07:12,150 --> 00:07:19,150 an attribute. It's supervised because it uses a class. Add an attribute, and AddClassfication. 60 00:07:20,110 --> 00:07:25,890 Here I get to choose in the configuration panel, the machine learning scheme. I'm going 61 00:07:25,890 --> 00:07:34,290 to choose J48, of course, and I'm going to outputClassification—put that True. 62 00:07:34,290 --> 00:07:39,300 That's configured it, and I'm going to apply it. It will add a new attribute. It's done it, 63 00:07:39,460 --> 00:07:46,010 and this attribute is the classification according to J48. Weka is very powerful. You can do 64 00:07:46,010 --> 00:07:52,510 all sorts of things with classifiers and filters. That's the end of the first class. 65 00:07:52,850 --> 00:07:58,930 There's a section of the book on Visualization. Please go and do the activity associated with this 66 00:07:58,930 --> 00:08:05,930 lesson, and I'll see you in the next class. Bye!