Computational Graph in Stock Marketing

These are going to be digital marketing agencies in Pune I can give them kind of almost garbage names. So I'm just going to say finals actually in what order. Yeah, we're turning finals opening and volumes so we're going to say finals openings and volumes is going to be equal to these the results of our stock data. And now we want to feed some of these particular finals and openings into our calculate price differences function. So we're going to have the end result be something like differences be equal to calculate price differences by feeding into finals and openings.

​

And then we can print out all of these price differences at the end of the day so we can just actually we don't even need to store the results here. We could just print them right out. So why don't we do that? So with this print statement of online advertising agency ready to go, we can give this a save and give it a run through okay and let's see if we get what we want digital marketing agencies in Pune to be printed out. Now keep in mind moving forward is that because we're subtracting the final price from the opening price a positive value store will mean that the stock price of the stock has gone up which means we probably want to buy at the end of the day and sell at the beginning of the next day because this means again that the price of the stock the next day is more expensive.

​

So we would actually make money on that. So that's why you'll see some positive and some negative values here. So if we actually compare these two values found in Chelsea species we want to test for accuracy. So we'll see. Take the first point. The final price or the other the opening price of the next day is just a Lysle and twelve hundred and eighty-four and ended up the previous day has about so should be about almost dollars change there. And indeed we do we get almost five years digital marketing agencies in Pune so that's accurate I guess. Yeah actually definitely is closer to than for.

​

Importance of Computational Graph

So how about the next day. Well, it looks like the price went up again by about little over $. So yes we got. again the next day it went up again by quite a lot actually. So it looks like our list of digital marketing agencies in Pune is actually pretty accurate so far. Just to verify our fifth data point shows a negative. So let's just check to see if the actually did go down and indeed it looks like it has it went from that would be this one here down to. Eighty-six points nine. So that's almost a. Almost. Yeah, a little over $ increase or decrease. And that's what we get about.

​

So I think we can conclude that this list of digital marketing agencies in Pune is accurate and that we now have a way to calculate all of the price differences that we've won. Remember our model is aimed at predicting what this price difference is going to be based on the volume exchange. And now we have all of the information we need to actually train and test our model. All that's left to do is build it and then of course train and test it. So that's where we can start working on in the section coming right up I'll just probably add a couple of comments here and then we can move on to this next section which is going to be actually building up our computation across from scratch.

​

All right guys now that we have a way to gain access to and format all the data that we need to feed digital marketing agencies in Pune into our model let's actually work on building that model up itself using some of our tents flow nodes. Now, this point I'm assuming that you know at least a little bit about tent's flow and about building a pretty simple regression-type model. That's what we're going to be building it's not going to be a hugely complex convolutional neural network or anything like that. So we'll model off the basic y equals W X plus B formats. So just try to follow along here I'm going to try not to go too far and explain what's going on along the way but it would help if you have had at least a little bit of insight into the tensile flow basics.

​

So I'd start off with we can actually get rid of all of this. We're not going to need it anymore. Zoom right in. So the construction of our actual graph again. ASTA y equals x plus B. OK. So I'd start off with Will we usually begin with our placeholder nodes so that's going to be our input into this is going to be a T.F. top placeholder and we'll give it a data type not too interested in the shape and the name. We can give that to. We don't really need to go for the data type we'll go with the T.F. to offload Tousey to and provide a name. This can just be X here. OK, so again this is going to act as a way to input data into our role models for training and testing purposes as well as just general evaluation. OK so next up is going to be w that's going to be a variable type node.