1 00:00:02,560 --> 00:00:03,660 ‫Hello everyone. 2 00:00:03,820 --> 00:00:10,530 ‫In this video we will learn how to split our available data and do best and green site 3 00:00:13,650 --> 00:00:23,500 ‫then we will train our more than on training set and we'll find out script on our test site to create 4 00:00:23,580 --> 00:00:24,960 ‫that screen is split. 5 00:00:24,970 --> 00:00:29,700 ‫First we need to import the function as string split from a scalar. 6 00:00:30,390 --> 00:00:34,060 ‫We will write from a scalar dot model selection 7 00:00:43,640 --> 00:00:44,430 ‫import 8 00:00:48,050 --> 00:00:49,610 ‫Green s split 9 00:00:58,470 --> 00:01:00,750 ‫known to use this function. 10 00:01:00,750 --> 00:01:04,470 ‫We first need to define our four variables. 11 00:01:04,470 --> 00:01:10,900 ‫That is our independent Green variable a word independent best variable. 12 00:01:10,920 --> 00:01:16,640 ‫Then number dependent green variable and then dependent test variable. 13 00:01:17,040 --> 00:01:24,210 ‫So we'll cleared this for variable we'll write X underscored Green 14 00:01:28,110 --> 00:01:29,600 ‫we'll write X and the score. 15 00:01:29,600 --> 00:01:30,030 ‫Best 16 00:01:34,420 --> 00:01:35,700 ‫wire and best score green 17 00:01:38,880 --> 00:01:41,520 ‫and Y underscore test. 18 00:01:41,990 --> 00:01:51,930 ‫These are our four data frames and this will get the value of the output of our screen test split functions. 19 00:01:51,970 --> 00:01:55,420 ‫We'll write Green underscore best 20 00:02:03,080 --> 00:02:09,200 ‫and in record will first mention the word independent variable which is x Maruti 21 00:02:13,460 --> 00:02:22,190 ‫then our dependent variable which is why Mighty Few remember we created this very well known last lecture. 22 00:02:22,190 --> 00:02:29,760 ‫Then the next parameter is a size as mentioned in order to re lecture. 23 00:02:29,970 --> 00:02:34,790 ‫We are splitting our data and it between data issue. 24 00:02:34,980 --> 00:02:42,740 ‫So what 80 percent of data will go into training set and no 20 percent of data will go in to test site. 25 00:02:42,900 --> 00:02:51,910 ‫That's right here we will mention zero point two this zero point two means 20 percent of data will go 26 00:02:51,910 --> 00:02:58,570 ‫into slice and the last parameter is that domestic. 27 00:02:58,660 --> 00:02:59,780 ‫Does that I know them. 28 00:02:59,790 --> 00:03:08,100 ‫But you can give any integer value we are providing this number just to get the same sample every time. 29 00:03:08,610 --> 00:03:16,700 ‫So if I mentioned random I say equal to one every time while running this test string slight every time 30 00:03:16,930 --> 00:03:20,830 ‫my training and test site will remain same. 31 00:03:20,830 --> 00:03:28,180 ‫So even if you are using the same random estate as we are using you will also get the same training 32 00:03:28,180 --> 00:03:29,160 ‫site. 33 00:03:29,560 --> 00:03:37,740 ‫So we will just mention random status CEDAW if you want to get the same training set. 34 00:03:38,490 --> 00:03:41,070 ‫You should also mention random search as 0 35 00:03:44,050 --> 00:03:53,870 ‫not just to check the number of rows and columns in our training and test set will right bring X underscore 36 00:03:53,870 --> 00:04:00,480 ‫Crane dot shape dot shape but you give me the number of rows and columns 37 00:04:20,520 --> 00:04:26,250 ‫now you can see 80 percent of our data is in training say so. 38 00:04:26,470 --> 00:04:39,760 ‫Four hundred and four observations are in our training set and rest of 1 0 2 observations are in asset. 39 00:04:39,900 --> 00:04:48,130 ‫Now we will follow the standard process of creating a linear regression model will first create an object. 40 00:04:48,210 --> 00:04:49,650 ‫This time we will name it. 41 00:04:49,680 --> 00:04:51,150 ‫ELAM underscored a 42 00:04:56,640 --> 00:05:10,070 ‫and will equate it to linear regression. 43 00:05:10,100 --> 00:05:16,040 ‫Now we will train our model from our training set that is extreme. 44 00:05:16,060 --> 00:05:17,960 ‫And by train right 45 00:05:21,030 --> 00:05:25,260 ‫Lemond scored a dog fit. 46 00:05:25,500 --> 00:05:39,980 ‫Then we will mention our training set which is x and the school train and Y underscore underscoring. 47 00:05:40,070 --> 00:05:46,390 ‫This statement will fit our model on our training set. 48 00:05:46,610 --> 00:05:52,360 ‫Now let's create the predicted value of y using this model. 49 00:05:53,240 --> 00:05:55,670 ‫So we will write y underscored test 50 00:06:00,000 --> 00:06:07,770 ‫underscored a equate to a limb a predict 51 00:06:12,730 --> 00:06:13,130 ‫here. 52 00:06:13,130 --> 00:06:20,830 ‫I'm predicting my test dependent variable so I will give my test independent variable so I will write 53 00:06:23,840 --> 00:06:25,340 ‫X and the score test. 54 00:06:28,970 --> 00:06:29,780 ‫If I run this 55 00:06:33,460 --> 00:06:40,270 ‫I have my predicted value of test set and by underscore best underscore a similarly we will create by 56 00:06:40,300 --> 00:06:46,450 ‫underscore Crane underscore to a will get the predicted values of our training set 57 00:06:55,830 --> 00:06:56,330 ‫this time. 58 00:06:56,340 --> 00:06:58,830 ‫We will use X and the school crane 59 00:07:06,080 --> 00:07:11,770 ‫now to check the artist squared value for our training and test data. 60 00:07:12,080 --> 00:07:13,820 ‫We will import another function 61 00:07:16,650 --> 00:07:20,410 ‫we will write from a skill and not my tricks. 62 00:07:20,440 --> 00:07:21,010 ‫Import 63 00:07:31,030 --> 00:07:33,670 ‫import are to that score score scored 64 00:07:37,800 --> 00:07:40,910 ‫you don't have to learn on this in Texas. 65 00:07:40,980 --> 00:07:46,040 ‫You can just save a copy of this notebook or you can search online. 66 00:07:46,080 --> 00:07:53,190 ‫The index are readily available online now to get the artist good value. 67 00:07:53,270 --> 00:07:58,100 ‫We just need to mention ah to underscore discord. 68 00:07:58,130 --> 00:07:59,910 ‫This is a function we imported 69 00:08:02,640 --> 00:08:06,230 ‫if we want to get more detail about this function. 70 00:08:06,270 --> 00:08:13,130 ‫We can just certitude and help by using question mark operator can just write. 71 00:08:13,130 --> 00:08:19,550 ‫Question mark and if we executed we will get all the details. 72 00:08:19,700 --> 00:08:20,800 ‫So here. 73 00:08:20,870 --> 00:08:30,970 ‫If you see in the same index we need to mention our y underscore crew which is the order is not Y values 74 00:08:31,450 --> 00:08:36,820 ‫then we need to mention the Y predicted values really do the same. 75 00:08:37,070 --> 00:08:42,730 ‫And just blows this right out to that school to. 76 00:08:46,090 --> 00:08:50,090 ‫And then record will first rate y underscore 77 00:08:54,370 --> 00:08:54,670 ‫this 78 00:09:00,120 --> 00:09:06,450 ‫so y underscore test is our or is another value and Y underscore best underscore a is the predicted 79 00:09:06,450 --> 00:09:07,810 ‫value of y variable. 80 00:09:07,900 --> 00:09:09,270 ‫So we'll just run this 81 00:09:12,720 --> 00:09:22,070 ‫so the is good value is zero point five for now let's call the artist good value for our training set 82 00:09:35,100 --> 00:09:40,220 ‫fit on this the artist good value for our training set is zero points and five. 83 00:09:41,710 --> 00:09:49,180 ‫You can also see that the artist good value for our test data is less than the artist good value for 84 00:09:49,180 --> 00:09:59,030 ‫our training site as we discuss in the world to be lectures that test artist good value is of more importance 85 00:09:59,120 --> 00:10:06,350 ‫as compared to training site and we should always look at our test score instead of creating artist 86 00:10:06,390 --> 00:10:10,300 ‫good values to evaluate the performance of our model 87 00:10:13,080 --> 00:10:17,850 ‫that so you split your data into distant dream and by 10.