Top Digital media agencies in Mumbai 2109
And that's because we know that the price of the stock is generally going to be different at the end of the day versus Digital media agencies in Mumbai the beginning of the next day. Our goal is to try and figure out whether that difference is positive or negative. So whether or not the price of the stock will go up the next morning or it will go down the next morning.
So if all model based on the volume at the end of the day will predict that the stock price will go up then we probably want to sell it right away the next morning because that will result in a net profit for us whereas digital marketing services if we know the stock price is going to go down or at least that's our model, predicts we probably want to hold onto that stock and not sell it otherwise we'd be losing money.
Now as I said this isn't going to be a super complex model which means it's probably not even going to be close to percent accuracy. In fact, the best results I've got using a pretty simple model is a little over percent. I think I've reached around percent with some of the stocks. Now you might be Digital media agencies in Mumbai looking at that number and saying well that's a garbage number percent accurate is not very good.
And you'd be absolutely right. With regards to and machine learning models generally, if we're not even close to percent accurate that's not a good model Astle we want very close to percent accurate or high s. Now in the case with these stock prediction models, I think the very best out there claims to be around percent. But that's a bold claim that just because there are so many factors that go into predicting stocks it's not just based on Passions and numerical data but there are so many other things of world events that affect the prices of stocks.
And that's just to name a few factors. So really it's an almost impossible task to get even close to percent accuracy with any stock prediction model no matter how complex you make it. The crux of it all is that the more data you feed in the more things the model has to consider. And so it might become confused past a certain point. That's why anything over percent accuracy in around percent Digital media agencies in Mumbai accuracy really isn't that bad. And honestly, if you were to do this long time with a percent accuracy result then we would probably result in net profits over the long term again.
Now that being said I don't claim responsibility for any money bet in this rather using this particular model. Nor do I even recommend you go about using this particular model with only percent accuracy and investing a lot of money into the hopes that this is going to get you rich. The purpose of this tutorial is mostly just to get you into the realm of making stock prediction models and of Rausch process that we would take to build something like this.
So my goal isn't to get close to percent accuracy. It's mostly just to show you the process of building up this model. With regards to stock market prediction. So if at the end of this tutorial you decide you want to use this model. I want to try and make some money using a percent accuracy then I'm not going to discourage you from doing so but I'm certainly not going to encourage you to rely on the percent accuracy figure.
This is more used for a learning tool and the grounds for a or the basis for a better model. OK but with that being said we have a little bit of background into what we're going to be building so let's talk about the order Digital media agencies in Mumbai in which we can implement our tasks. So we'll start by first we'll just explore our data sets. I think we'll start off with just one stock, to begin with. I think we'll use the gold exchange stock.
So we'll just go to be using and predicting gold prices that we can incorporate some more towards the end and see what kind of different results we get based on those other stocks. OK. So once we're comfortable with how the day is going to be represented We'll take a look at how to import format and then manipulate that retrieve data. And normally this is probably going to be the majority of the tutorials just doing this. Like I said the model will actually be building is pretty simple and it doesn't take a long time to train the model either.
OK. So we're going to need to perform a lot of manipulations to make sure that data can be fed accurately and precisely into our model and we're going to get real results. So after we are comfortable with how the data has been manipulated we can build and train our computational graphs like a said is Digital media agencies in Mumbai going to be pretty simple and won't require any kind of like complicated neural networks or anything like that just because we are again only relying on this one fact which has volume after we have the computational graph built and we can train it.
Then we're going to test the accuracy of our results. So we'll be able to build some kind of an accuracy function that's going to tell us our model will probably end up being around percent accurate give or take a little bit depending on which stock we're using and when we're going to be testing. So, by the way, I'll show you where it's again the data sets that we're going to be using but I've also provided those data sets as C S V files in the Resources folder.
Anyway, so I recommend you use those ones just that we're on the same page. So the very final result at the end our model will be able to take in the volume exchange for a particular day. Okay. And that will predict whether or not the stock price will increase or decrease the next morning as a result. Okay. So now you have a brief overview as to what we're going to be covering in this project Digital media agencies in Mumbai and how we're going to go about doing it. So we can jump right into some actual coding but of course, we'll start by exploring the data sets first.
I'll show you where to get up. What website will go to be extracting the data from? And then what the CSP sheets will look like. Okay. Now moving through this tutorial some or all of the videos will feature me with kind of a husky voice I'm Lysle so I do apologize if that's bothering you at some point in time. I'll try and keep it short so you guys don't have to listen to it too much. OK. So let's get started in the next section by exploring our datasets.