WEBVTT

00:00.080 --> 00:02.480
Hello developers, and welcome back.

00:02.520 --> 00:09.120
In this section we're going to build an amazing application called House Price Prediction Application.

00:09.120 --> 00:14.640
We're going to mix and integrate machine learning with Android.

00:14.800 --> 00:22.840
We'll train a simple model with TensorFlow from scratch to learn the equation and correlation between

00:23.120 --> 00:27.240
the size of house and its price.

00:27.240 --> 00:35.480
So we're going to start with installing TensorFlow, then generating random data, then reshaping data,

00:35.640 --> 00:43.400
building TensorFlow model, training the model, get the trained weights and bias plotting the results.

00:43.400 --> 00:50.480
So we're going to see the coloration and the actual data and the TensorFlow prediction.

00:50.680 --> 00:57.200
We're going to evaluate the model by plotting the errors and see the model loss.

00:57.360 --> 01:00.640
Also we're going to save and export the model.

01:00.680 --> 01:05.840
How to export our model in TensorFlow format.

01:05.840 --> 01:11.830
Also we're going to learn how to load models, how to load the exported models.

01:11.950 --> 01:20.150
Also, the most important thing we're going to learn is exporting models as TF Lite for Android TF Lite

01:20.150 --> 01:24.750
format optimized for mobile and edge devices like microcontrollers.

01:24.750 --> 01:31.470
So we're going to export our model and use it in Android Studio.

01:31.630 --> 01:37.750
So through this amazing tutorial we're going to build this application.

01:38.070 --> 01:49.150
So this application let me show you what it does I gave the model the TensorFlow model 100 data 100

01:49.390 --> 01:55.790
sizes of the the the houses and their prices.

01:56.070 --> 02:03.710
It predicted the price of any any house you gave us as a size.

02:03.710 --> 02:05.110
So give me the size.

02:05.110 --> 02:06.710
I'll give you the price.

02:06.950 --> 02:10.150
This is the TensorFlow job okay.

02:10.350 --> 02:16.210
So here I'll give it a house size in thousands of square feet.

02:16.210 --> 02:21.530
So 1.5 means 1500ft².

02:21.690 --> 02:24.690
And let me predict this.

02:24.890 --> 02:29.090
So here click on predict house price.

02:29.090 --> 02:33.490
And the price of the house is 6.548.

02:33.570 --> 02:45.410
The model learned the relationship when when we we give when we gave the model 100 data 100 size and

02:45.410 --> 02:46.250
the price.

02:46.250 --> 02:49.170
There is an equation learned by the model.

02:49.370 --> 02:54.930
And this learned equation is used to predict the price of the house.

02:55.050 --> 03:01.530
For example, 1.5 if we go to here to the, uh, the house prediction using TensorFlow, this graph

03:01.570 --> 03:05.930
1.5, it's approximately 6.5 okay.

03:05.970 --> 03:11.090
So if we make a prediction so it's approximately is 6.5.

03:11.090 --> 03:15.050
So it predicted right numbers.

03:15.250 --> 03:15.730
Okay.

03:15.890 --> 03:22.750
So we're going to make sure that our model is well trained and everything is working fine.

03:22.750 --> 03:26.230
Let me try another prediction example.

03:26.230 --> 03:36.430
So predict house price and with 2000ft² house, its price would be $8,000.

03:36.510 --> 03:39.790
If we make a projection here we see that it is.

03:40.030 --> 03:44.670
It's close to $8.05 thousand.

03:44.710 --> 03:45.190
Okay.

03:45.350 --> 03:49.670
So congratulations guys for reaching this section.

03:49.670 --> 03:52.910
Also, this is a big section.

03:53.110 --> 04:00.190
And um, I want from you to understand everything I want from you to follow me.

04:00.390 --> 04:09.190
There are a lot a lot of notes I want from you to write them down, because they help you to memorize

04:09.190 --> 04:12.430
and understand what we're going to build.

04:12.470 --> 04:15.750
Guys, this is a very important section.

04:16.030 --> 04:22.990
This is the building block of your machine learning and Android integration career.