Learn Stock Exchange by digital agencies Bangalore

Introduction Stock Exchange

So then we create some variables for them up here the number of data points we're going to be using we created this function that is going to be load stock data. So this is going to be used to basically load in the pieces of data from the CAC sheets that we one of the digital agencies Bangalore to get to load these three columns price open and volume price represents a final price at the end of the day opening the opening price at the beginning of the day and volume is the volume of stock exchange throw a particular day. So it's formatted them a bit and Drone getting as much data as we specified in these number of data points here. Okay. So next up was a call to get that last crucial piece of information that was the calculate price differences function.

Session 1

So we basically take a final price of the stock at the end of the day and get the difference between that and the opening price. The next day so points opening price and then store those differences in the Saran. Now we're just returning it in this function case that's entirely what this function is used to. Probably digital agencies Bangalore should have added some comments to this. I think I will at the end just in the final source code that you do have those comments there but otherwise. Well, actually we'll skip over this fanatic's we did implement that just yet. Our next stop was building Al computational graph we modeled it off to the simple y equals W X plus B in which x is a placeholder where we impose volumes WMP with variables.

Session 2

That our model is going to modify and why is going to be the final outputs we'll be able to answer digital agencies Bangalore in some volumes and then get the price difference for those corresponding volumes. He also had this predict to note the scourge we use to compare online marketing company actual versus predicted values. Next up is our last function at our optimizer. And these in order to train our model we ran in this session here for the certain number of times a number of epochs I think we set to. Learning right was point and basically just running this optimize over and over again allowing it to change and B until we get the desired values for these. And our model is as accurate as it can be.

Session 3

After this we tested actually I guess we parse it out with data to just to show you guys that existed but this wasn't a crucial step. This is just full of visual aid. So after that, we've finished off by couch lasing the cure. We built this A.C. function which measures how many digital agencies Bangalore guesses our model got correct or what it should be. I'm guessing so. Correct. If the guess value is negative and the actual values are negative. I call that a win whereas if there was some discrepancy that was lost in that case OK. And then we just overturning the accuracy about percent for gold too because the spigot will change depending on which stock you're evaluating and which Assent you're feeding into your model to both train and test.

Session 4

And then we just finished off here by adding in the rest of the CSFB sheets. OK. So some suggestions to improve this project or to go a bit further is to test these guys out download some extra CCSB sheets and test and train your model with those making sure that you're changing this. Maybe try modifying the learning rate of digital agencies Bangalore and the number of parks to see if that changes how accurate your model ends up being. Another thing you could take into consideration if you want to build a more complex model would be to add a neural network with several different layers as well as to maybe factor in the final price and see if that also coupled with volume has an effect on the opening price at the end of the day.

So this is just a few suggestions. Again my job here was just to show you the basics and build a really simple model for you guys to basically just show you the process how we would go about building something like this and then leave it open for you guys to improve upon. And as always this isn't terribly accurate. Please don't rely on this to try to get rich because the stock markets are ultimately kind of unpredictable and there's a lot of different things that could happen that would cause them to cause prices to go up and down.

So honestly any model is never going to be percent accurate percent reliable. OK. But with that being said I'll leave you guys here if you want to try those suggestions. Absolutely that's fantastic. If you want to move on to the next tutorials that are cool too. But I just want to say big thank you to digital agencies Bangalore for watching these tutorials really appreciate the support. Hope you guys can learn a lot from them least take something out with these tutorials. I'm excited to see what you guys can build from them. All right. Thanks for watching and we'll see you next time.