WEBVTT 00:01.030 --> 00:06.820 Now that they have learned about all the algorithms which are present in supervised learning, we will 00:06.820 --> 00:15.910 discuss about pipeline generation Pipeline basically helps us to create a complete flow from the region 00:16.180 --> 00:23.710 to testing of data so that we can do this without writing the piece of code again and again. 00:25.770 --> 00:33.000 So a machine learning pipeline is used to help automate machine learning flows, they operate by enabling 00:33.000 --> 00:40.680 a sequence of data to be transformed and correlated together in a model that can be tested and evaluated 00:40.680 --> 00:44.170 to achieve an outcome, whether positive or negative. 00:44.850 --> 00:47.490 It helps us in saving more and more. 00:47.560 --> 00:55.050 This can be loaded from a file using a biplane, and it also helps business applications to plug and 00:55.050 --> 00:56.720 connect the pipeline directly. 00:58.840 --> 01:03.080 So for this, let us have a look at the pipeline structure. 01:03.310 --> 01:08.980 So basically a pipeline will help us to automate the Indian machine. 01:08.980 --> 01:10.030 Learning will flow. 01:10.040 --> 01:17.140 And using a pipeline, we can simply create a pipeline, begin to create the entire flow, how the details 01:17.140 --> 01:19.990 should be going to the transformation phase. 01:20.230 --> 01:26.890 And we will see the modalities so that next time when we want to use the model, we can simply pick 01:26.890 --> 01:34.150 the fight from the pipeline and use the pipe itself and simply get the output from it. 01:34.270 --> 01:36.730 We don't have to write the codes again and again. 01:38.370 --> 01:39.820 So this is the pipeline. 01:40.110 --> 01:46.860 So in the pipeline, we have the training data which goes into this complete pipeline, which we have, 01:46.860 --> 01:48.330 this is the entire pipeline. 01:48.750 --> 01:53.880 So in the pipeline for the day that will be cleaned, then it will be normalized. 01:54.280 --> 01:56.520 Then we will impute the value. 01:57.530 --> 02:04.910 Then feature engineering will be done after that, we can apply any of these or any of the model which 02:04.910 --> 02:11.360 we want to apply in, for example, random photos or decision three, any type of model we can delete 02:11.360 --> 02:11.690 here. 02:12.140 --> 02:18.380 And once we have generated the model and despite this computer, we can simply put the testing data 02:18.380 --> 02:23.690 inside the pipeline and make the prediction so we don't have to write this code again and again or train 02:23.690 --> 02:24.920 the model again and again. 02:25.130 --> 02:31.640 We can simply pick this pipeline and place it anywhere and use it so we will learn about how we can 02:31.640 --> 02:38.450 implement in the next session and see how this can actually be implemented.