1 00:00:00,000 --> 00:00:04,000 So, guys, uh, we are going to continue the discussion with respect to this particular project. 2 00:00:04,000 --> 00:00:07,000 Uh, you can see over here all the libraries has been installed. 3 00:00:07,000 --> 00:00:08,000 Okay. 4 00:00:08,000 --> 00:00:10,000 And I even did not get any kind of error. 5 00:00:10,000 --> 00:00:11,000 Okay. 6 00:00:11,000 --> 00:00:14,000 Now, I have also activated this specific environment in my terminal. 7 00:00:14,000 --> 00:00:20,000 So let me just, uh, hide my terminal right now, because I will not be requiring what I will do. 8 00:00:20,000 --> 00:00:23,000 Uh, along with this one last thing, I also need to do the installation. 9 00:00:23,000 --> 00:00:25,000 That is nothing but ipykernel. 10 00:00:25,000 --> 00:00:31,000 So since the first experiment that I'm actually going to do, I'm going to do in my Jupyter notebook. 11 00:00:31,000 --> 00:00:37,000 And for that I require this ipykernel library so that I'll be able to attach a kernel and execute my 12 00:00:37,000 --> 00:00:38,000 Jupyter notebook file. 13 00:00:38,000 --> 00:00:43,000 So here I'm actually going to write pip install Ipykernel. 14 00:00:43,000 --> 00:00:43,000 Okay. 15 00:00:43,000 --> 00:00:50,000 So once I execute this you will be able to see that the installation will take place. 16 00:00:50,000 --> 00:00:50,000 Okay. 17 00:00:50,000 --> 00:00:57,000 So once this ipykernel has been executed, then, uh, I'm going to probably start my coding. 18 00:00:57,000 --> 00:01:04,000 And the first thing that I will do is that I will read this entire CSV file and let me do also one more 19 00:01:04,000 --> 00:01:04,000 thing. 20 00:01:04,000 --> 00:01:10,000 Let me just go ahead and create my experiments dot Ipynb file. 21 00:01:10,000 --> 00:01:10,000 Okay. 22 00:01:10,000 --> 00:01:13,000 So I'm going to use this IPO Nvl file. 23 00:01:13,000 --> 00:01:16,000 And in this my entire experiment will happen okay. 24 00:01:16,000 --> 00:01:20,000 So right now here you can see my Ipy kernel library is basically getting installed. 25 00:01:20,000 --> 00:01:22,000 That is going to take some amount of time. 26 00:01:22,000 --> 00:01:23,000 Now it is done. 27 00:01:23,000 --> 00:01:26,000 Now I will quickly, uh, hide this terminal. 28 00:01:26,000 --> 00:01:27,000 Okay. 29 00:01:28,000 --> 00:01:30,000 Um, let me just go ahead and select my kernel. 30 00:01:30,000 --> 00:01:34,000 First of all, like here I have selected my kernel with, uh, Python 3.11. 31 00:01:34,000 --> 00:01:38,000 Okay, now I will just go ahead and make some more code cells. 32 00:01:38,000 --> 00:01:41,000 Now let's start this entire coding okay. 33 00:01:41,000 --> 00:01:47,000 So first of all, uh, what I'm actually going to do is that, uh, I'm going to import pandas. 34 00:01:47,000 --> 00:01:50,000 So let me just go ahead and import pandas as PD. 35 00:01:51,000 --> 00:01:58,000 Uh, along with this I will go ahead and import from sklearn dot model selection because I also have 36 00:01:58,000 --> 00:01:59,000 to do the train test split. 37 00:01:59,000 --> 00:02:00,000 Right. 38 00:02:00,000 --> 00:02:05,000 So for that I will be using this model selection import train test split. 39 00:02:05,000 --> 00:02:09,000 I hope you know this from the machine learning which we have actually used. 40 00:02:10,000 --> 00:02:16,000 And then along with this I will also go ahead and import from sklearn dot pre-processing. 41 00:02:16,000 --> 00:02:21,000 So here I will just go ahead and write pre processing. 42 00:02:22,000 --> 00:02:24,000 And then I'm going to import two important libraries. 43 00:02:24,000 --> 00:02:30,000 One is standard scalar and the other one that I'm actually going to use is something called as label 44 00:02:30,000 --> 00:02:31,000 encoder. 45 00:02:32,000 --> 00:02:32,000 Okay. 46 00:02:32,000 --> 00:02:37,000 So this two libraries I'm actually going to use along with this I'm going to go ahead and import pickle 47 00:02:37,000 --> 00:02:38,000 file okay. 48 00:02:38,000 --> 00:02:42,000 Pickle is specifically for pickling because when we are using the standard scalar and label encoder 49 00:02:42,000 --> 00:02:47,000 it is important that we go ahead and pickle this file so that we can reuse it when we are doing the 50 00:02:47,000 --> 00:02:48,000 deployment. 51 00:02:48,000 --> 00:02:52,000 So here are some of the libraries that we have already done the installation. 52 00:02:52,000 --> 00:02:54,000 And we are just going to import it. 53 00:02:54,000 --> 00:02:54,000 Okay. 54 00:02:55,000 --> 00:02:56,000 So these are some of the basic libraries. 55 00:02:56,000 --> 00:03:00,000 Now let us go ahead and load the data set okay. 56 00:03:00,000 --> 00:03:06,000 So for this I will say data is equal to PD dot read underscore CSV. 57 00:03:06,000 --> 00:03:13,000 And here I'm going to basically call churn underscore modeling dot csv file. 58 00:03:13,000 --> 00:03:19,000 Now over here you'll be able to see I'll just go ahead and display data dot head. 59 00:03:19,000 --> 00:03:25,000 Okay, so these are all my values or all my features that are available over here. 60 00:03:25,000 --> 00:03:28,000 I've already told you in my previous problem statement what we are going to do. 61 00:03:28,000 --> 00:03:34,000 We are going to take this credit score, geography, gender, age, tenure, balance, number of product, 62 00:03:34,000 --> 00:03:39,000 has credit card, is active member and estimated salary as my independent feature. 63 00:03:39,000 --> 00:03:43,000 And I'm going to predict whether the person is going to exit the bank or not. 64 00:03:43,000 --> 00:03:43,000 Okay. 65 00:03:43,000 --> 00:03:47,000 And this is the past data from the bank which shows which person has exited or not. 66 00:03:47,000 --> 00:03:49,000 So this is just the top five records. 67 00:03:49,000 --> 00:03:55,000 Now the first important thing is that this first three features are not that important, right? 68 00:03:55,000 --> 00:03:58,000 You have row number, you have customer ID, you have surname. 69 00:03:58,000 --> 00:04:01,000 This definitely will not play a very important role. 70 00:04:01,000 --> 00:04:10,000 So first of all what we will do we will go ahead and pre-process the data. 71 00:04:10,000 --> 00:04:10,000 Okay. 72 00:04:10,000 --> 00:04:17,000 What we are going to do here, we are just going to drop irrelevant features okay. 73 00:04:17,000 --> 00:04:20,000 Which we can directly see it okay. 74 00:04:20,000 --> 00:04:22,000 So here uh I will use this data variable. 75 00:04:22,000 --> 00:04:27,000 And from this I will say hey go ahead and drop which all features we need to drop okay. 76 00:04:27,000 --> 00:04:29,000 So drop is a function over here. 77 00:04:29,000 --> 00:04:32,000 We we really need to drop this row number. 78 00:04:32,000 --> 00:04:36,000 So I'll copy and paste it over here. 79 00:04:36,000 --> 00:04:41,000 Then we have something called as customer ID. 80 00:04:43,000 --> 00:04:43,000 Okay. 81 00:04:44,000 --> 00:04:47,000 Then we have something called as surname. 82 00:04:48,000 --> 00:04:48,000 Okay. 83 00:04:48,000 --> 00:04:51,000 And we are going to drop all these things with axis equal to one. 84 00:04:51,000 --> 00:04:53,000 Axis is equal to one basically means column wise. 85 00:04:53,000 --> 00:04:54,000 Okay. 86 00:04:54,000 --> 00:04:59,000 So once we do this particular drop we will be able to get the data value. 87 00:04:59,000 --> 00:05:00,000 So here is my data value. 88 00:05:00,000 --> 00:05:06,000 Now here you can see my features are starting from credit score geography gender age balance and all 89 00:05:06,000 --> 00:05:08,000 these values are there okay perfect. 90 00:05:08,000 --> 00:05:09,000 So this is good enough. 91 00:05:09,000 --> 00:05:10,000 Uh we have done this. 92 00:05:10,000 --> 00:05:17,000 Now here you can actually see we have a column called as, uh, geography over here. 93 00:05:17,000 --> 00:05:17,000 Right. 94 00:05:17,000 --> 00:05:18,000 So geography is there. 95 00:05:18,000 --> 00:05:19,000 Okay. 96 00:05:19,000 --> 00:05:22,000 And then, um, we also have this gender. 97 00:05:22,000 --> 00:05:25,000 We have this, uh, this two features. 98 00:05:25,000 --> 00:05:27,000 We specifically have, uh, geography and gender. 99 00:05:28,000 --> 00:05:33,000 Now geography and gender, uh, is specifically in this particular case, you can see it is a categorical 100 00:05:33,000 --> 00:05:34,000 variable. 101 00:05:34,000 --> 00:05:34,000 Okay. 102 00:05:34,000 --> 00:05:41,000 Now in the case of categorical variable, what I can actually do is that I can apply some kind of encoding 103 00:05:41,000 --> 00:05:42,000 for this categorical variable. 104 00:05:42,000 --> 00:05:42,000 Okay. 105 00:05:42,000 --> 00:05:44,000 And that is what I am actually going to do. 106 00:05:45,000 --> 00:05:51,000 Now let's go ahead and encode one categorical variable. 107 00:05:53,000 --> 00:06:01,000 Now in this I will say hey um let's take first of all gender I'll say label underscore encoder underscore 108 00:06:01,000 --> 00:06:10,000 gender is equal to label encoder okay I'm going to just use this and I'm going to say hey DF of gender 109 00:06:10,000 --> 00:06:16,000 or I'll say data of gender because we have used this particular feature, data of gender. 110 00:06:16,000 --> 00:06:20,000 And I'll just use this label encoder. 111 00:06:22,000 --> 00:06:25,000 Encoder underscore gender dot fit. 112 00:06:26,000 --> 00:06:29,000 And along with fit we will also be using transform. 113 00:06:29,000 --> 00:06:32,000 And we have already discussed this why we use fit underscore transform. 114 00:06:32,000 --> 00:06:35,000 And here I'm going to take this particular data. 115 00:06:35,000 --> 00:06:39,000 I'm going to basically use apply this on my gender itself. 116 00:06:39,000 --> 00:06:39,000 Right. 117 00:06:39,000 --> 00:06:41,000 So now let's go ahead and see the data. 118 00:06:41,000 --> 00:06:44,000 And let's see with respect to this particular gender. 119 00:06:44,000 --> 00:06:48,000 So here you can see okay I'm getting an error the spelling mistake I'll go ahead and write this one 120 00:06:48,000 --> 00:06:49,000 okay. 121 00:06:49,000 --> 00:06:54,000 So once I do this so my gender will have two main features right. 122 00:06:54,000 --> 00:06:55,000 One is male and female. 123 00:06:55,000 --> 00:06:59,000 So that male and female with the help of label encoder has got converted to zeros and ones. 124 00:06:59,000 --> 00:07:02,000 Okay, we have also seen this in machine learning okay. 125 00:07:02,000 --> 00:07:10,000 So this is one of the encoding techniques with respect to the categorical variable okay I hope you are 126 00:07:10,000 --> 00:07:11,000 able to understand this. 127 00:07:12,000 --> 00:07:16,000 So guys uh over here in the gender we saw that our values are male and female. 128 00:07:16,000 --> 00:07:19,000 And so that is the reason we use label encoder. 129 00:07:19,000 --> 00:07:21,000 And we changed it to zeros and ones. 130 00:07:21,000 --> 00:07:23,000 But what about the geography column? 131 00:07:23,000 --> 00:07:26,000 In this geography column you basically have 2 to 3 features right. 132 00:07:26,000 --> 00:07:30,000 2 to 3 categories like France, Spain and Japan. 133 00:07:30,000 --> 00:07:31,000 Germany. 134 00:07:31,000 --> 00:07:31,000 Right. 135 00:07:31,000 --> 00:07:33,000 So all this country names are specifically there. 136 00:07:33,000 --> 00:07:42,000 So here if we apply, just imagine that if we go ahead and uh, apply, uh, you know the label encoder. 137 00:07:42,000 --> 00:07:47,000 So let me just go ahead and open my epi pen so that I will explain you over here. 138 00:07:47,000 --> 00:07:52,000 Why, uh, why I'm specifically focus on, uh, on this particular geographic column. 139 00:07:52,000 --> 00:07:52,000 Right. 140 00:07:52,000 --> 00:07:58,000 So let's say if I just go ahead and convert this into a categorical column. 141 00:07:58,000 --> 00:08:03,000 So here you'll be able to see I will be able to get France as zero, Spain as one, and Germany as two. 142 00:08:03,000 --> 00:08:09,000 And when these types of numbers are specifically coming up, then there is a problem because, uh, 143 00:08:09,000 --> 00:08:16,000 when we assign a value of Germany to two, it the at the end of the day, an is all about, you know, 144 00:08:16,000 --> 00:08:17,000 numerical calculations. 145 00:08:18,000 --> 00:08:22,000 Since the number or since the label of Germany is two, it will just consider that Germany is greater 146 00:08:22,000 --> 00:08:25,000 than Spain or Spain is greater than France. 147 00:08:25,000 --> 00:08:25,000 Right. 148 00:08:25,000 --> 00:08:27,000 So this should not happen. 149 00:08:27,000 --> 00:08:33,000 So for this particular case, what we will do is that we will not directly convert this with the help 150 00:08:33,000 --> 00:08:35,000 of label encoder, but instead what we will do. 151 00:08:35,000 --> 00:08:39,000 We'll go ahead and write some code and we will use one hot encoding okay. 152 00:08:39,000 --> 00:08:43,000 Now one hot encoding will just actually give us some values with respect to zeros and ones. 153 00:08:43,000 --> 00:08:44,000 Okay. 154 00:08:44,000 --> 00:08:51,000 So now uh, I will just go ahead and write one hot encode geographic column. 155 00:08:51,000 --> 00:08:51,000 Okay. 156 00:08:51,000 --> 00:08:54,000 So here I'm just going to write geographic column over here. 157 00:08:54,000 --> 00:08:59,000 Now quickly let's first of all go ahead and import the one hot encoder. 158 00:08:59,000 --> 00:09:03,000 So I will say from sklearn dot pre-processing. 159 00:09:04,000 --> 00:09:13,000 From sklearn dot pre-processing I'm going to just import one hot one hot encoder. 160 00:09:13,000 --> 00:09:18,000 Let me just see this particular I'm seeing the documentation inside wise. 161 00:09:18,000 --> 00:09:22,000 So it will be helpful for me to also check it out directly instead of wasting time. 162 00:09:22,000 --> 00:09:23,000 Okay. 163 00:09:23,000 --> 00:09:25,000 And, uh, writing each and every thing. 164 00:09:25,000 --> 00:09:25,000 Okay. 165 00:09:25,000 --> 00:09:29,000 So here what I'm actually going to do, I'm going to specifically use one hot encoder. 166 00:09:29,000 --> 00:09:32,000 Now quickly I will just go ahead and use this one hot encoder. 167 00:09:32,000 --> 00:09:33,000 I will initialize this. 168 00:09:34,000 --> 00:09:37,000 And this time there is a parameter which is called as sparse. 169 00:09:37,000 --> 00:09:37,000 Right. 170 00:09:37,000 --> 00:09:41,000 So I will just make this feature sparse is equal to false. 171 00:09:41,000 --> 00:09:44,000 I will let you know why we are basically using this. 172 00:09:44,000 --> 00:09:47,000 And once you see the encoded values then you'll be able to understand. 173 00:09:48,000 --> 00:09:54,000 I'll create a variable called as one hot underscore encoder underscore geo okay. 174 00:09:54,000 --> 00:09:58,000 So this will basically be my values over here okay. 175 00:09:58,000 --> 00:10:06,000 Now uh quickly uh, the next step that uh we are going to specifically use is nothing but geo underscore 176 00:10:06,000 --> 00:10:07,000 encoded value. 177 00:10:07,000 --> 00:10:19,000 And I'll say, hey, let's apply this one hot encoder one hot encoder underscore geo dot fit dot fit 178 00:10:19,000 --> 00:10:29,000 under okay one port underscore geo dot fit underscore transform on which column basically on my geographic 179 00:10:29,000 --> 00:10:30,000 column. 180 00:10:30,000 --> 00:10:30,000 Right. 181 00:10:30,000 --> 00:10:36,000 So here I'm going to give my geography column right now. 182 00:10:36,000 --> 00:10:38,000 Once this is done you'll be able to see that. 183 00:10:38,000 --> 00:10:47,000 Let's go ahead and see how I'm actually going to get this geo underscore sorry geo underscore encoder. 184 00:10:48,000 --> 00:10:50,000 Um and let's execute this. 185 00:10:50,000 --> 00:10:54,000 So here it says uh got an unexpected argument. 186 00:10:54,000 --> 00:10:57,000 Sparse is equal to false okay. 187 00:10:57,000 --> 00:10:59,000 Let me just remove it for now. 188 00:10:59,000 --> 00:11:00,000 Onwards. 189 00:11:00,000 --> 00:11:00,000 Okay. 190 00:11:00,000 --> 00:11:02,000 Let's execute this. 191 00:11:02,000 --> 00:11:06,000 So here I'm actually able to get a sparse row matrix okay. 192 00:11:06,000 --> 00:11:11,000 Uh I think in one hot encoder I have something called as sparse. 193 00:11:11,000 --> 00:11:13,000 Sparse is equal to false. 194 00:11:13,000 --> 00:11:14,000 Let's see. 195 00:11:15,000 --> 00:11:19,000 Um, got an unexpected argument called as sparse. 196 00:11:19,000 --> 00:11:22,000 I think in the recent version the sparse is not present. 197 00:11:22,000 --> 00:11:24,000 Let me just go ahead and use directly this one. 198 00:11:24,000 --> 00:11:25,000 Okay. 199 00:11:25,000 --> 00:11:30,000 Now once I get this Jio encoder, uh, over here. 200 00:11:30,000 --> 00:11:31,000 Right. 201 00:11:31,000 --> 00:11:36,000 What I will do is that, uh, if I use this Jio encoder, okay. 202 00:11:36,000 --> 00:11:41,000 Or if I also go ahead and see what all features specifically we have. 203 00:11:41,000 --> 00:11:45,000 If I try to convert this into a numpy array, I will be able to see all the values. 204 00:11:45,000 --> 00:11:54,000 So first of all what I will do I'll just go ahead and write geo or sorry one hot one hot underscore 205 00:11:54,000 --> 00:12:00,000 encoder underscore geo dot get underscore features. 206 00:12:00,000 --> 00:12:03,000 Underscore names underscore out. 207 00:12:03,000 --> 00:12:04,000 Right. 208 00:12:04,000 --> 00:12:07,000 So if I go ahead and see this parameter okay. 209 00:12:07,000 --> 00:12:08,000 Sorry out. 210 00:12:08,000 --> 00:12:11,000 And I here I'm just going to give my function. 211 00:12:11,000 --> 00:12:17,000 And inside this I'm just going to give my V column geography. 212 00:12:17,000 --> 00:12:22,000 So if I go ahead and execute it it says get underscore features okay. 213 00:12:22,000 --> 00:12:29,000 It should be feature sorry get underscore feature underscore names underscore out okay. 214 00:12:30,000 --> 00:12:35,000 So here you'll be able to see with respect to the geography I have this three features that is getting 215 00:12:35,000 --> 00:12:38,000 converted into a one hot encoder. 216 00:12:38,000 --> 00:12:39,000 Right. 217 00:12:39,000 --> 00:12:44,000 So, uh, with respect to this, if I try to convert, I'll use this particular features and I will 218 00:12:44,000 --> 00:12:49,000 be using this, uh, sparse matrix that I've actually got in the form of data frame. 219 00:12:49,000 --> 00:12:53,000 Now, in order to get that in the form of data frame, what I will do, I will just go ahead and write 220 00:12:53,000 --> 00:12:59,000 geo underscore encoder or geo underscore encoder. 221 00:12:59,000 --> 00:13:06,000 Uh, here I will use this particular value and I will convert this entire into PD dot data frame. 222 00:13:06,000 --> 00:13:08,000 So I'll just go ahead and write PD dot data frame. 223 00:13:08,000 --> 00:13:12,000 And here what I'll say this is my data G0 underscore encoder. 224 00:13:12,000 --> 00:13:18,000 Along with that uh if I really want to assign the columns to this, it will be nothing but this entire 225 00:13:18,000 --> 00:13:18,000 values. 226 00:13:18,000 --> 00:13:22,000 Because from this I'm actually able to get the columns okay. 227 00:13:23,000 --> 00:13:26,000 so this is my entire, uh, data frame. 228 00:13:26,000 --> 00:13:30,000 And this is what I'm actually getting as my encoded underscore DF. 229 00:13:30,000 --> 00:13:30,000 Okay. 230 00:13:30,000 --> 00:13:32,000 So let me just go ahead and execute this. 231 00:13:32,000 --> 00:13:39,000 So once I execute this and if I go ahead and see my g underscore encoded underscore df I'm getting an 232 00:13:39,000 --> 00:13:39,000 error. 233 00:13:39,000 --> 00:13:45,000 Uh, the reason is uh shape of value uh is 1000 comma one. 234 00:13:45,000 --> 00:13:49,000 So guys here you can see that uh I'm actually getting this particular error. 235 00:13:49,000 --> 00:13:50,000 You know, it. 236 00:13:50,000 --> 00:13:54,000 Uh, shape of value that is passed is 10,000 comma one. 237 00:13:54,000 --> 00:13:59,000 And it is specifically implying that it needs 10,000 comma three because here, uh, we are giving three 238 00:13:59,000 --> 00:13:59,000 columns. 239 00:13:59,000 --> 00:14:00,000 Right. 240 00:14:00,000 --> 00:14:03,000 Now why this specific problem is happening. 241 00:14:03,000 --> 00:14:05,000 See if I go ahead and print this geo encoder. 242 00:14:05,000 --> 00:14:09,000 It shows compressed sparse row matrix of dtype right. 243 00:14:09,000 --> 00:14:16,000 Now if I probably convert this geo encoder to two underscore array okay. 244 00:14:16,000 --> 00:14:19,000 And if I execute it uh two underscore array. 245 00:14:19,000 --> 00:14:19,000 Sorry. 246 00:14:19,000 --> 00:14:20,000 It should be two array. 247 00:14:20,000 --> 00:14:24,000 So here you will be able to see that I am able to convert this into a sparse value. 248 00:14:24,000 --> 00:14:25,000 Right. 249 00:14:25,000 --> 00:14:30,000 And uh that is the reason, you know before I was just giving this instead of giving this, I will just 250 00:14:30,000 --> 00:14:31,000 go ahead and give this. 251 00:14:31,000 --> 00:14:32,000 Okay. 252 00:14:32,000 --> 00:14:38,000 So here you will be able to see that if I go ahead and give it now And if I go ahead and see this encoded 253 00:14:38,000 --> 00:14:44,000 underscore f uh df so here you can see that I'm actually able to get this geography France, geography, 254 00:14:44,000 --> 00:14:48,000 Germany and Spain, wherever France is that value would be one and remaining. 255 00:14:48,000 --> 00:14:50,000 All will be zero wherever Germany is. 256 00:14:50,000 --> 00:14:54,000 Um, that will be one like over here and wherever Spain is, that will be one. 257 00:14:54,000 --> 00:14:54,000 Okay. 258 00:14:54,000 --> 00:14:57,000 So this is how you fix this particular problem. 259 00:14:57,000 --> 00:15:03,000 Again, I'm not going to probably, uh, delete any of this or, uh, remove any or part of this particular 260 00:15:03,000 --> 00:15:09,000 video because I want everyone to see the errors that we specifically face, you know, because once 261 00:15:09,000 --> 00:15:15,000 you see the error, you probably become much more confident in, uh, solving that particular error. 262 00:15:15,000 --> 00:15:15,000 Right. 263 00:15:15,000 --> 00:15:18,000 So let's go ahead and combine all the columns now. 264 00:15:18,000 --> 00:15:23,000 So here I will just go ahead and combine all the columns. 265 00:15:23,000 --> 00:15:30,000 Call the one hot encoded columns with the original data. 266 00:15:31,000 --> 00:15:34,000 Okay, with the original data. 267 00:15:34,000 --> 00:15:41,000 So here what I'm actually now going to do is that quickly, uh, and this is really important, uh, 268 00:15:41,000 --> 00:15:45,000 because without this trust me, because the new part of the code. 269 00:15:45,000 --> 00:15:45,000 Right. 270 00:15:45,000 --> 00:15:48,000 What we did with respect to transformation, we really need to update that. 271 00:15:48,000 --> 00:15:50,000 So I will just go ahead and update it. 272 00:15:50,000 --> 00:15:54,000 And I will write data is equal to PD dot concatenation. 273 00:15:54,000 --> 00:16:04,000 And here first of all I have to concatenate the new the new geo encoded de underscore DF with our this 274 00:16:04,000 --> 00:16:05,000 particular data. 275 00:16:05,000 --> 00:16:06,000 Right. 276 00:16:06,000 --> 00:16:09,000 So first of all I will just go ahead and do the data dot drop. 277 00:16:09,000 --> 00:16:11,000 So I will go ahead and write data dot drop. 278 00:16:11,000 --> 00:16:16,000 And here I'm going to use my column which is called as geography because I don't require it any time 279 00:16:16,000 --> 00:16:16,000 now. 280 00:16:16,000 --> 00:16:20,000 And here I'm going to just use it as access is equal to one okay. 281 00:16:21,000 --> 00:16:26,000 So once this is dropped the next thing is that I will just go ahead and concatenate with my geo underscore 282 00:16:26,000 --> 00:16:28,000 encoded underscore df. 283 00:16:28,000 --> 00:16:30,000 And this time I'm just going to give my access equal to one. 284 00:16:30,000 --> 00:16:30,000 Right. 285 00:16:30,000 --> 00:16:36,000 So now if I go ahead and see my data dot head you'll be seeing that okay gender has become zero one. 286 00:16:36,000 --> 00:16:42,000 And along with this there'll be three more features like geography, France, Germany and Spain that 287 00:16:42,000 --> 00:16:43,000 has got concatenated. 288 00:16:43,000 --> 00:16:51,000 Okay, so this is how we have actually converted our categorical features into numerical features. 289 00:16:51,000 --> 00:16:57,000 And uh, we have basically used two important type of uh encoding techniques. 290 00:16:57,000 --> 00:17:00,000 One is label encoder and the other one is uh one hot encoder. 291 00:17:02,000 --> 00:17:03,000 So or so I hope. 292 00:17:03,000 --> 00:17:11,000 Uh, you have, uh, you are able to understand this entire thing, uh, and, uh, see, everything 293 00:17:11,000 --> 00:17:16,000 I'm trying to do step by step will do it very much in a slower way so that you get all the points, 294 00:17:16,000 --> 00:17:23,000 you understand things, and, uh, you go in, you see, uh, all this, ah, like, all machine learning 295 00:17:23,000 --> 00:17:24,000 kind of feature engineering. 296 00:17:24,000 --> 00:17:25,000 Only we did nothing. 297 00:17:25,000 --> 00:17:27,000 Something new would not apply. 298 00:17:27,000 --> 00:17:27,000 Over here. 299 00:17:27,000 --> 00:17:31,000 Everything is same, like how we have actually performed right now. 300 00:17:31,000 --> 00:17:39,000 Uh, let's quickly go ahead and, uh, see, uh, one is what I have used, uh, label encoding one, 301 00:17:39,000 --> 00:17:40,000 I have used one hot encoder. 302 00:17:40,000 --> 00:17:41,000 Right. 303 00:17:41,000 --> 00:17:46,000 So I have to probably go ahead and save all this particular files right in the form of pickle file, 304 00:17:46,000 --> 00:17:52,000 because this same one hot encoding and same label encoding I'm actually going to use in the later stages. 305 00:17:52,000 --> 00:17:53,000 Right. 306 00:17:53,000 --> 00:17:56,000 So label encoder gender I will try to go ahead and save it. 307 00:17:56,000 --> 00:18:00,000 And then one hot encoder with respect to this G0 I'm going to save it okay. 308 00:18:00,000 --> 00:18:02,000 So let's go ahead and save this particular file in the pickle. 309 00:18:02,000 --> 00:18:07,000 So that in our end to end project we have to probably use this. 310 00:18:07,000 --> 00:18:15,000 So I'll just go ahead and write save the encoders and scalar Okay. 311 00:18:15,000 --> 00:18:17,000 So I'll say hey with open. 312 00:18:17,000 --> 00:18:18,000 I'll open a pickle file. 313 00:18:18,000 --> 00:18:22,000 I'll say label underscore encoder. 314 00:18:23,000 --> 00:18:24,000 Okay. 315 00:18:24,000 --> 00:18:25,000 Label underscore encoder. 316 00:18:25,000 --> 00:18:28,000 Underscore geo okay. 317 00:18:29,000 --> 00:18:31,000 Label underscore encoder underscore gender. 318 00:18:31,000 --> 00:18:34,000 First of all we'll go ahead and save this pickle file. 319 00:18:34,000 --> 00:18:40,000 And for this I have already imported my pickle file I'll go ahead and open this in my write byte mode, 320 00:18:40,000 --> 00:18:44,000 because this is a deserialized or serialized file itself. 321 00:18:44,000 --> 00:18:50,000 So we will be writing in the form of bytes and this as file I have created a temporary variable. 322 00:18:50,000 --> 00:18:57,000 I'll just go ahead and write pickle dot dump, and I will dump this entire label encoder gender inside 323 00:18:57,000 --> 00:18:59,000 this particular file okay. 324 00:18:59,000 --> 00:19:07,000 Similarly you'll also be seeing that I will go ahead and open with open, uh, one more pickle file. 325 00:19:07,000 --> 00:19:08,000 I'll just go ahead and create it. 326 00:19:08,000 --> 00:19:13,000 That is nothing but one hot encoder underscore geo dot pickle file. 327 00:19:14,000 --> 00:19:19,000 And this will also get opened in my write byte mode okay WB mode. 328 00:19:19,000 --> 00:19:22,000 And here I can see as file okay. 329 00:19:22,000 --> 00:19:26,000 And this also we will go ahead and dump it. 330 00:19:27,000 --> 00:19:30,000 So I'll say pickle dot dump. 331 00:19:30,000 --> 00:19:36,000 And here also I'm going to use my one hot. 332 00:19:38,000 --> 00:19:41,000 Encoder underscore Geo. 333 00:19:41,000 --> 00:19:45,000 And finally I'm just going to convert this into my file okay. 334 00:19:46,000 --> 00:19:48,000 So let's go ahead and open it. 335 00:19:48,000 --> 00:19:52,000 So I'll say with open or okay this two file right now I'll save it. 336 00:19:52,000 --> 00:19:57,000 So once I've save it uh you can see this is my pickle file that has got created. 337 00:19:58,000 --> 00:20:05,000 So till here uh the basic transformation that I have actually done right, that all has been completed. 338 00:20:06,000 --> 00:20:09,000 And, uh, we have almost done this kind of transformation, and we have saved it. 339 00:20:09,000 --> 00:20:12,000 Okay, now, uh, let's do one thing. 340 00:20:12,000 --> 00:20:14,000 Okay, so you have done the transformation. 341 00:20:14,000 --> 00:20:17,000 Now let's quickly divide our data. 342 00:20:17,000 --> 00:20:25,000 Divide the data set into independent and dependent features. 343 00:20:28,000 --> 00:20:33,000 Now you know that my my data dot head. 344 00:20:33,000 --> 00:20:35,000 If I just go ahead and execute it. 345 00:20:35,000 --> 00:20:35,000 Right. 346 00:20:35,000 --> 00:20:40,000 So this is data dot head here. 347 00:20:40,000 --> 00:20:47,000 Exit is my exit is my column which I will be keeping as my dependent feature and remaining all this 348 00:20:47,000 --> 00:20:48,000 particular column. 349 00:20:48,000 --> 00:20:49,000 I will keep it in my independent features. 350 00:20:49,000 --> 00:20:50,000 Right. 351 00:20:50,000 --> 00:20:53,000 So quickly let's go ahead and do that. 352 00:20:53,000 --> 00:20:58,000 So I will just try to go ahead and divide my data set into independent and dependent features. 353 00:20:58,000 --> 00:20:59,000 So I'll say data dot drop. 354 00:21:00,000 --> 00:21:08,000 And here I will go ahead and write exited comma access is equal to one okay y is equal to data. 355 00:21:09,000 --> 00:21:12,000 And here I will just go ahead and say exited right. 356 00:21:12,000 --> 00:21:17,000 So this will basically be my dependent Y will be my dependent X will be my independent feature. 357 00:21:17,000 --> 00:21:28,000 Then we will split the data in training when training and testing sets Okay. 358 00:21:29,000 --> 00:21:38,000 Now let me just go ahead and write x train comma x test comma y train comma y test quickly. 359 00:21:38,000 --> 00:21:43,000 And then this is very much common because we are using it from machine learning x comma y. 360 00:21:43,000 --> 00:21:46,000 Uh let's go ahead and use some test size. 361 00:21:46,000 --> 00:21:48,000 So test size will be 0.2. 362 00:21:49,000 --> 00:21:53,000 And some random underscore state will be 42. 363 00:21:55,000 --> 00:21:58,000 So here we have also what we need to do. 364 00:21:58,000 --> 00:21:59,000 One more thing guys. 365 00:21:59,000 --> 00:22:01,000 We'll scale down this features. 366 00:22:01,000 --> 00:22:01,000 Okay. 367 00:22:01,000 --> 00:22:02,000 Scale this features. 368 00:22:02,000 --> 00:22:06,000 And the scaling up of this all the features. 369 00:22:06,000 --> 00:22:08,000 Also we will be specifically doing okay. 370 00:22:08,000 --> 00:22:14,000 Now scaling we will use the scaler scaler standard scalar okay. 371 00:22:14,000 --> 00:22:17,000 And uh once we'll go ahead and initialize it. 372 00:22:17,000 --> 00:22:27,000 Uh, here I'm just going to write X train is equal to scalar dot fit transform fit transform on my uh 373 00:22:27,000 --> 00:22:28,000 the entire X train data. 374 00:22:28,000 --> 00:22:34,000 And for my X test underscore data X underscore test data I'm just going to use scalar dot transform. 375 00:22:35,000 --> 00:22:38,000 And here I'm just going to go ahead and write x underscore test. 376 00:22:39,000 --> 00:22:44,000 So this transformation is specifically required I will quickly go ahead and execute it. 377 00:22:46,000 --> 00:22:47,000 So. 378 00:22:50,000 --> 00:22:50,000 Okay. 379 00:22:50,000 --> 00:22:54,000 Um x comma y okay. 380 00:22:54,000 --> 00:22:55,000 Perfect. 381 00:22:55,000 --> 00:22:57,000 So let's go ahead and execute this. 382 00:22:57,000 --> 00:23:03,000 So here you'll be able to see this will be my X train data which will be the uh it will be basically 383 00:23:03,000 --> 00:23:06,000 scaling is getting uh, getting applied with the help of standard scalar. 384 00:23:06,000 --> 00:23:08,000 And similarly I have my X test. 385 00:23:09,000 --> 00:23:14,000 So once this is done, we'll also go ahead and save this scalar file with the help of uh in the form 386 00:23:14,000 --> 00:23:15,000 of pickle format. 387 00:23:15,000 --> 00:23:15,000 Okay. 388 00:23:15,000 --> 00:23:18,000 We will be requiring all this pickle format. 389 00:23:18,000 --> 00:23:23,000 So with open I will say hey go ahead and use my scalar dot pickle. 390 00:23:23,000 --> 00:23:29,000 And uh it will be in the right byte mode Okay. 391 00:23:29,000 --> 00:23:31,000 Uh, right byte mode. 392 00:23:33,000 --> 00:23:38,000 Now, once we are using this right byte mode, I will just go ahead and create my context over here 393 00:23:38,000 --> 00:23:39,000 as file. 394 00:23:39,000 --> 00:23:47,000 And we'll go ahead and dump this dump this scalar dot scalar comma file okay. 395 00:23:47,000 --> 00:23:49,000 Whatever scalar value is there. 396 00:23:49,000 --> 00:23:51,000 So let's go ahead and execute it. 397 00:23:51,000 --> 00:23:53,000 So here you have each and every thing. 398 00:23:53,000 --> 00:23:54,000 Here you have label encoder pickle. 399 00:23:54,000 --> 00:23:55,000 You have one hot encoder pickle. 400 00:23:55,000 --> 00:23:56,000 We have scalar pickle. 401 00:23:56,000 --> 00:23:57,000 This is perfect. 402 00:23:58,000 --> 00:24:02,000 So in short in this video what I have done is that I've just cleaned some data set. 403 00:24:02,000 --> 00:24:06,000 I've applied all type of uh feature engineering that I want to do. 404 00:24:06,000 --> 00:24:06,000 Okay. 405 00:24:06,000 --> 00:24:09,000 I have divided this into dependent and independent features. 406 00:24:09,000 --> 00:24:14,000 What all, uh, main feature engineering methods that we have applied is nothing, but we have applied 407 00:24:14,000 --> 00:24:17,000 both our standard scalar. 408 00:24:17,000 --> 00:24:22,000 We have applied, uh, we have converted categorical features into numerical features using label encoder, 409 00:24:22,000 --> 00:24:24,000 one hot encoder, everything we have specifically done. 410 00:24:25,000 --> 00:24:28,000 So, uh, now my data is specifically ready. 411 00:24:28,000 --> 00:24:28,000 Okay. 412 00:24:28,000 --> 00:24:29,000 This is my data. 413 00:24:30,000 --> 00:24:34,000 Now on this particular data, I'm going to just go ahead and train my artificial neural network. 414 00:24:34,000 --> 00:24:38,000 And that is what I'm actually going to show you in the next video. 415 00:24:38,000 --> 00:24:40,000 So I hope you like this particular video. 416 00:24:40,000 --> 00:24:41,000 Uh, I will see you in the next video. 417 00:24:41,000 --> 00:24:42,000 Thank you.