WEBVTT 00:00.080 --> 00:02.800 Hello developers, and welcome back. 00:02.840 --> 00:09.720 In this section we're going to build this amazing app called Car Mileage Predictor App. 00:10.080 --> 00:13.280 This app is a complex app. 00:13.520 --> 00:23.320 It's an advanced application that uses machine learning to predict the mileage based on many parameters. 00:23.360 --> 00:29.440 The first step we're going to get the data from and from an online source. 00:29.440 --> 00:32.480 And by the way, those are real data. 00:32.480 --> 00:34.880 So we're going to work with real data. 00:34.920 --> 00:39.400 I'm going to show you the real data and dealing with real data. 00:39.400 --> 00:45.480 So this would be very exciting dealing with with real data. 00:45.680 --> 00:49.840 So our app will accept in the input layer cylinders. 00:49.840 --> 00:50.800 Displacement. 00:50.800 --> 00:51.720 Horsepower. 00:51.760 --> 00:52.360 Weight. 00:52.400 --> 00:53.440 Acceleration. 00:53.440 --> 00:56.960 Model year origin one, origin two and origin three. 00:57.200 --> 01:02.760 Those are the nine parameters we're going to insert in the input layer. 01:03.040 --> 01:10.260 Then we're going to move to create the hidden layer number one with 64 neurons, then the hidden layer 01:10.260 --> 01:19.020 number two with 64 neurons, and getting the output layer, which is the mileage and predicting the 01:19.100 --> 01:20.100 mileage. 01:20.100 --> 01:25.020 So it's not an easy and simple application. 01:25.220 --> 01:29.100 It's gonna be a little bit complex and advanced. 01:29.140 --> 01:34.700 We're going to use those data in order to train our model with Google Colab. 01:34.700 --> 01:42.420 So the learning objectives is are data loading and exploration, data pre-processing and cleaning, 01:42.420 --> 01:49.420 building a neural network with TensorFlow, Keras model training and evaluation, and making predictions 01:49.620 --> 01:52.140 throughout the whole process. 01:52.140 --> 01:56.060 We're going to use libraries in Python. 01:56.140 --> 02:02.740 We're going to display the data the distribution of mpg in histograms. 02:02.780 --> 02:05.860 Also we're going to learn about heat maps. 02:06.100 --> 02:11.640 We're going to learn about the data parameters and the layers. 02:11.920 --> 02:18.720 Also, we're going to train the model getting the evaluation and much, much more. 02:18.760 --> 02:27.640 Also, we're going to make the difference and a comparison between the actual versus predicted mpg. 02:28.000 --> 02:28.520 Okay. 02:28.760 --> 02:33.320 So it's gonna be a very exciting application. 02:33.320 --> 02:36.840 We're gonna build together using the real data. 02:36.840 --> 02:46.040 And we're going to transform those data into a data frames that our model accept. 02:46.080 --> 02:54.360 Also going to train our model in order to start predicting the data according to this chart or or similar 02:54.360 --> 02:55.400 to this chart. 02:55.600 --> 02:59.440 Also we're going to integrate it with Android Studio. 02:59.440 --> 03:01.760 So this is our application. 03:01.960 --> 03:05.080 For example the user enters four cylinders. 03:05.320 --> 03:09.000 Displacement 150 cc, horsepower 100. 03:09.080 --> 03:12.230 Weight Thousand pounds. 03:12.270 --> 03:15.550 Acceleration zero point the 15. 03:15.590 --> 03:18.950 It's from zero to 60s model year. 03:18.990 --> 03:23.430 Between 70 and 82 and origin would be America. 03:23.430 --> 03:26.350 If I click predict mpg it predicts. 03:26.390 --> 03:37.070 This number if I select European predict mpg and it's 21.39 select Japanese predict mpg. 03:37.310 --> 03:38.670 Which is 22. 03:38.990 --> 03:43.390 So also we're going to handle the errors. 03:43.390 --> 03:47.310 So if the user like exceeded the. 03:47.350 --> 03:51.030 Model year for example like this predict mpg. 03:51.150 --> 03:52.670 We get the error. 03:52.710 --> 03:55.390 Please fix the model year. 03:55.430 --> 03:57.710 Should be between 70 and 82. 03:57.830 --> 04:03.150 So all of those things we're gonna learn in our application. 04:03.190 --> 04:05.070 Our lovely application. 04:05.070 --> 04:09.830 And by the way it's not an easy application. 04:09.830 --> 04:11.710 It's hard application. 04:11.750 --> 04:20.850 Advanced application that we're going to use machine learning AI, and we're going to get real data 04:20.890 --> 04:22.010 based on real data. 04:22.010 --> 04:27.290 We're gonna predict our new data and the MPG. 04:27.330 --> 04:27.850 Okay. 04:28.210 --> 04:31.330 So this is a very exciting application. 04:31.610 --> 04:41.090 This is a very, very interesting and very advanced application that the TensorFlow is used to build 04:41.090 --> 04:42.850 machine learning models. 04:43.010 --> 04:50.570 Also, we're going to use Android Studio and how to integrate machine learning with our Kotlin Jetpack 04:50.570 --> 04:52.090 Compose application. 04:52.130 --> 04:58.130 Again guys, this is a very interesting application we're gonna build together. 04:58.130 --> 05:03.370 So I am very excited to start building this application. 05:03.530 --> 05:11.050 Please support us with rating five stars, to start building more and more applications and support 05:11.050 --> 05:13.210 us making new courses.