WEBVTT 00:02.070 --> 00:06.060 Hello and welcome to this tutorial in previously told he'll be out. 00:06.190 --> 00:11.700 Understood activation function in this tutorial with the loan court function. 00:11.880 --> 00:15.750 Let us begin the court function majors. 00:15.800 --> 00:23.760 The performance of machine learning model for a given day to day court function quantifies the error 00:23.760 --> 00:26.790 between predicted, valued and expected values. 00:27.000 --> 00:32.160 And after that, it presents the result in the form of a single real number. 00:33.030 --> 00:38.070 So we use the cost function to major performance of a machine learning model. 00:41.280 --> 00:46.260 Let us understand some notations that we use in the equation of cost function. 00:47.820 --> 00:49.880 Why is the neurons to value? 00:50.070 --> 00:53.340 Whereas VI had is the neurons predicted value. 00:56.420 --> 00:59.330 This is the equation of quadratic court function. 00:59.500 --> 01:03.440 C is equal to summation of Y minus Y had square. 01:03.680 --> 01:11.030 And this is divided by and the errors are more prominent in this cost function. 01:11.210 --> 01:19.820 And this is because the squared off Y minus VI had and this calculation of the cost function can cause 01:20.030 --> 01:23.480 a slowdown in learning speed of machine learning model. 01:24.290 --> 01:27.290 So this is all about decode cost function. 01:30.410 --> 01:35.410 As you can see here, this is the equation for cross and proper cost function. 01:36.890 --> 01:43.450 Due to this equation, the cross entropy cost function allows faster learning of machine learning more 01:45.100 --> 01:47.630 due to that equation of cost function. 01:47.660 --> 01:52.680 We can say that the larger the difference between VI and VI had faster. 01:52.790 --> 01:55.320 The neuron learns that needs faster. 01:55.380 --> 01:57.290 The machine learning model loans. 01:57.980 --> 02:04.430 So we can say that the cross entropy cost function is better than the code rettig cost function. 02:07.580 --> 02:14.430 No, we can say that there are two key aspects with which a neural network loans that is activation 02:14.430 --> 02:16.560 function and cost function. 02:18.160 --> 02:20.550 We are still missing a key step here. 02:20.790 --> 02:24.560 That is the actual learning process of the neural network. 02:26.250 --> 02:33.480 Now we have to figure out a way so we can use the cost function and neurons to obtain the correct prediction 02:34.020 --> 02:35.010 in simple words. 02:35.040 --> 02:40.290 We can say that we have to figure out the actual learning process of neural network. 02:41.070 --> 02:43.660 We will do that in next tutorial bill. 02:43.720 --> 02:45.300 Then happy learning.