1 00:00:00,870 --> 00:00:08,580 Harlette continue with our project, and this video will do as Kelly, so in this project, will you 2 00:00:08,950 --> 00:00:15,480 make it hard to get on the scale that are in that range from zero and one because computer only understand 3 00:00:15,480 --> 00:00:16,250 zero and one. 4 00:00:16,770 --> 00:00:19,470 So how do the formula work? 5 00:00:20,220 --> 00:00:22,300 Let's try some formula for that. 6 00:00:22,770 --> 00:00:23,720 It's very easy. 7 00:00:23,740 --> 00:00:24,180 So. 8 00:00:26,640 --> 00:00:27,180 Scale. 9 00:00:28,270 --> 00:00:30,760 Equal X minus. 10 00:00:32,580 --> 00:00:33,010 Mean. 11 00:00:35,040 --> 00:00:36,780 So that total. 12 00:00:39,260 --> 00:00:40,460 We'll be revived by. 13 00:00:42,240 --> 00:00:43,980 X marks. 14 00:00:45,040 --> 00:00:47,170 Minor, I mean. 15 00:00:49,060 --> 00:00:57,310 And does it do to perform features, scowling, we can use the prepossessing package available in the 16 00:00:57,310 --> 00:01:03,610 second library so that I could learn library is a free software machine learning library for a python 17 00:01:03,610 --> 00:01:04,560 programming language. 18 00:01:04,900 --> 00:01:12,360 So it's free chose BRIAREOS classification Chryson and clustering algorithms, including the Support 19 00:01:12,380 --> 00:01:23,740 Whitemarsh, SVM, Random Forest Radio and boosting games and dibby dibby debates scan and is SDI to 20 00:01:24,520 --> 00:01:31,360 incorporate with the python numerical and scientific libraries, nimbies and Skype. 21 00:01:32,240 --> 00:01:38,260 So remember to import our library that is not present in the initial distribution of Python. 22 00:01:38,560 --> 00:01:39,730 You must be in. 23 00:01:40,220 --> 00:01:41,920 Come on, come on in here. 24 00:01:41,920 --> 00:01:49,010 We don't need to install anything because everything is installed in a Google CoLab, so we just Yoda, 25 00:01:49,300 --> 00:01:58,330 Escalon, not prepossessing Buckeye's Provisor or Common Utility Functions and Transformers Kos's to 26 00:01:58,330 --> 00:02:04,120 modify the features available in the representation that bessus our need. 27 00:02:04,840 --> 00:02:09,970 So we need to import our baksh is very easy so that. 28 00:02:12,920 --> 00:02:18,800 Creation and not from České Lord, not pray. 29 00:02:20,050 --> 00:02:20,890 Processing. 30 00:02:23,030 --> 00:02:23,690 Import. 31 00:02:25,500 --> 00:02:25,960 Mean. 32 00:02:27,750 --> 00:02:28,260 Max. 33 00:02:29,820 --> 00:02:30,630 And. 34 00:02:32,080 --> 00:02:32,830 Skalla. 35 00:02:34,340 --> 00:02:41,120 So to scale features between people and minimum and maximum value, in our case between zero and one, 36 00:02:41,540 --> 00:02:48,650 so that the maximum absolute value of a scale to the units, I mean, my scale of function. 37 00:02:49,460 --> 00:02:53,360 So let's start by defining the scale object. 38 00:02:55,040 --> 00:02:56,990 So that really scared the object. 39 00:02:58,460 --> 00:03:02,450 Scale are equal men. 40 00:03:03,650 --> 00:03:04,670 Mark Skalla. 41 00:03:05,720 --> 00:03:09,930 And now you're to have a confirmation of what we are doing. 42 00:03:10,250 --> 00:03:14,750 We need to bring the perimeter that we will use for the next resizing. 43 00:03:16,390 --> 00:03:17,470 So bring. 44 00:03:18,410 --> 00:03:19,700 Skalla. 45 00:03:20,930 --> 00:03:21,680 Dafydd. 46 00:03:22,530 --> 00:03:23,160 Dadar. 47 00:03:24,920 --> 00:03:31,850 And then run the sale and we got Muscala equal feature featurette from Zero and one. 48 00:03:33,160 --> 00:03:40,510 So the fit method can boost the minimum and maximum that is to be used for the later Skellig. 49 00:03:41,610 --> 00:03:42,630 So that riot. 50 00:03:45,190 --> 00:03:50,340 They can put the minimum. 51 00:03:52,010 --> 00:03:53,810 And maximum. 52 00:03:55,050 --> 00:04:01,410 That is to be you for later scaling. 53 00:04:04,900 --> 00:04:05,580 So. 54 00:04:07,170 --> 00:04:17,790 Now, what we were doing, we knew the scale of we just so let ground near Cazale and the variable where 55 00:04:17,790 --> 00:04:19,050 you data scale. 56 00:04:20,700 --> 00:04:24,240 Equal Skyler Dafydd. 57 00:04:25,990 --> 00:04:27,400 Underscored transform. 58 00:04:29,010 --> 00:04:29,580 Dadar. 59 00:04:30,530 --> 00:04:37,460 And a freight train for my third fit to the data and then transform it so that run this El. 60 00:04:40,870 --> 00:04:48,800 So an umpire right now is retired and advisable to report and always in the starting from Vandersloot 61 00:04:48,820 --> 00:04:51,370 offramps or at least for comparison purposes. 62 00:04:52,330 --> 00:04:52,980 So, like. 63 00:04:55,360 --> 00:04:59,430 What does your data scale equa? 64 00:05:00,990 --> 00:05:04,380 PD or data frame. 65 00:05:06,460 --> 00:05:10,630 And then data scale and then column. 66 00:05:12,250 --> 00:05:13,000 Equa. 67 00:05:14,550 --> 00:05:15,960 Behesht names. 68 00:05:17,920 --> 00:05:25,930 And to verify that that information work area, we due to some basic statistics that we had already 69 00:05:25,930 --> 00:05:27,070 calculated so. 70 00:05:27,970 --> 00:05:30,040 That being so Samari. 71 00:05:31,660 --> 00:05:32,410 Equa. 72 00:05:33,430 --> 00:05:34,450 Detat scale. 73 00:05:37,190 --> 00:05:37,970 Distri. 74 00:05:44,270 --> 00:05:48,290 And then summary, Igor. 75 00:05:50,130 --> 00:05:51,030 Samori. 76 00:05:54,740 --> 00:05:56,240 Address, post. 77 00:05:57,860 --> 00:05:59,150 And then we need to bring. 78 00:06:00,340 --> 00:06:01,180 Summary. 79 00:06:04,280 --> 00:06:07,460 And we got our summary, so as you say that. 80 00:06:11,870 --> 00:06:18,020 Would that driver with a preference for our resolve, every variable is included in the range between 81 00:06:18,020 --> 00:06:20,900 zero and one, so as you say, is from zero and one. 82 00:06:23,010 --> 00:06:29,900 Now, all villages have values zero and one we will now move on to visually visual analysis. 83 00:06:30,390 --> 00:06:33,970 For example, what we can do is block the variable. 84 00:06:34,260 --> 00:06:42,090 So what is referred to as Whiskas charge is a graphic representation that is used to describe the distribution 85 00:06:42,090 --> 00:06:50,190 of assemble by symbol dispersion and boxes and indexes so I can be presented either horizontally or 86 00:06:50,190 --> 00:06:53,280 vertically by making a rectangle. 87 00:06:54,890 --> 00:07:04,310 Partition divided by two segments, so the rectangle box is delineated by the first quarter. 88 00:07:04,550 --> 00:07:12,860 Twenty five quanti the third quartile, 70 75 quanti and divided by the media and 50. 89 00:07:13,820 --> 00:07:17,960 Fifty by the S.I, so. 90 00:07:20,370 --> 00:07:25,620 The median is 50 percent high, so let's have some tax for that. 91 00:07:27,950 --> 00:07:32,030 The first one is lower, riska. 92 00:07:35,090 --> 00:07:38,200 And then a 75. 93 00:07:39,420 --> 00:07:40,350 By S.I. 94 00:07:41,480 --> 00:07:49,100 The next one is Maryann, with a 50 per cent high. 95 00:07:50,300 --> 00:07:51,470 The next one, A. 96 00:07:52,470 --> 00:07:53,520 Seven DFI. 97 00:07:54,490 --> 00:07:55,480 By Senti. 98 00:07:57,260 --> 00:07:59,530 And the next one will be. 99 00:08:02,080 --> 00:08:02,770 Upper. 100 00:08:04,460 --> 00:08:09,060 Whisker, and that is the end of this video. 101 00:08:09,470 --> 00:08:13,580 I hope you enjoy it and I will see you in the next video.