I Predicted the Apple Stock Prices with AI - You Won’t Believe the Results! 📈

hey everyone welcome back to the channel today we're diving deep into the world of stock price prediction specifically the Apple stock price now before we start just a quick disclaimer this video is for educational purposes only stock price prediction can be fun and insightful but it's not Financial advice always do your own research or consult a financial advisor before making any investment decisions now let's get into it first we need to set up our coding environment we're using some powerful python libraries today numpy pandas matplotlib tensorflow and scikit learn these will help us manipulate data visualize Trends and of course build our machine learning model once you've got these installed we import them into our script we're also pulling in historical data for Apple courtesy of the Y Finance Library let's get that data and take a quick peek at it so what we just did was fetch Apple stock data from 2010 all the way to the end of 2023 this includes every high low open close and volume of trade for each day but for our purposes we're only interested in the Clos prices next we move on to pre-processing our data we need to normalize our data to make it easier for the model to learn patterns think of it as leveling the playing field so that no one data point is too small or too large compared to the others by now our data has been normalized squeezed into a nice manageable range between 0 and 1 this step is crucial because it helps our model converge faster and better but hey don't worry we'll bring everything back to its original scale later here's where things get interesting we need to create sequences out of our data basically generating time windows that the model can learn from for example if today's stock price is X what was it for the last 10 days this helps the model understand Trends over time all right we've just created sequences of 10 days each our model will use the last 10 days to predict the price on day 11 and so on pretty cool right since we're focusing on the bigger picture in this video and not diving into every function or parameter if you're curious about the details like how specific libraries work or the nitty-gritty of model building be sure to check out my other machine learning videos on the channel they're packed with everything you need to become a pro now it's time to split our data into training and testing sets we'll use about 2/3 of our data for training and the remaining for testing this way we can check how well our model performs on data it hasn't seen before we've split the data and reshaped it to fit our lsdm model the lsdm needs the data in a specific shape and that's exactly what we've done here now it's ready for some machine learning magic let's talk about our model we're using an LS SDM or long short-term memory it is a type of recurrent neural network that's designed to remember important information over long sequences while forgetting the irrelevant Parts this makes it perfect for time series data like stock prices where understanding the trends and patterns over time is Key to making accurate predictions our model is now built it's got two lsdm layers with 50 units each followed by a couple of dense layers to Output our prediction we've compiled it with the atom Optimizer and means squared error loss function now let's get this model trained here's the Moment of Truth training the model we'll run it for 10 epics let's see how it performs so we trained our model for 10 epics you probably saw the accuracy improving as we progressed but remember accuracy isn't always the best metric for stock price prediction what matters more is how close our predictions are to the actual prices and for that we'll look at mean squared error MSE and mean absolute error ma now that our model is trained let's evaluate its performance on the test data this will give us an idea of how well it might perform in the real world all right our test loss and accuracy are out but what's more important here is the loss specifically the mean squared error which tells us how far off our predictions are on average time to see how well our model did by making some predictions and plotting them against the actual data visualization is key to understanding how close our models predictions are to reality we've just made our predictions and scaled them back to the original stock prices let's plot them and see how well our model performed finally let's plot everything out this will give us a visual comparison of our models predictions against the actual stock prices and there you have it the Blue Line represents the actual Apple stock prices the orange line is what our model predicted on the training set and the green line is the models predictions on the test set looks pretty close doesn't it all right folks that's a wrap on predicting Apple's stock prices using deep learning remember while our model looks pretty cool the stock market can be unpredictable so always trade carefully but hey if you think you can predict the next Apple stock surge why not leave your prediction in the comments maybe you'll be the next Warren Buffett or not but hey it's fun to guess thanks for watching don't forget to like subscribe and hit that notification Bell so you don't miss our next Deep dive until then Happy predicting and even happier investing