Stock Price Prediction | AI in Finance

Can AI be used in the financial sector? Of course! In fact, finance was one of the pioneering industries that started using AI in the early 80s for market prediction. Since then, major financial firms and hedge funds have adopted AI technologies for everything from portfolio optimization, to credit lending, to stock betting. In this video, we'll go over all the different ways AI can be used in applied finance, then build a stock price prediction algorithm in python using Keras and Tensorflow.

Code for this video:

Please Subscribe! And like. And comment. That's what keeps me going.

Want more education? Connect with me here:

More learning resources:

Join us in the Wizards Slack channel:

Sign up for the next course at The School of AI:

And please support me on Patreon:
Signup for my newsletter for exciting updates in the field of AI:

[DrMantisTobaggan]

Dude, You are awesome!!! I just graduated with a Masters In Data Analytics, and my program is no where near the level of your knowledge and the topics you teach. You are the best professor I have ever had in Data!

[jarvy]

I usually speed up videos to 1.25x to 1.5x speed; but I never need to do that with Siraj's videos!

[ssaurabh.]

This is the best video I have ever come across in AI. I tried many a times to learn ML and AI but almost everytime i left it in between because i was unaware the possibility and application of it, what can i do after learning it.. and here in your video without diverting a single sec, i can trace my roadmap for next 10 year. Thanks of it

[willc]

Dude, I just found you today, you are a GOD! I just finished a Data Science Course and I'm going to begin your "Learn Machine Learning in 3 Months" within the next week. I just wanted to stop by and say that you are amazing, and I appreciate your hard work in putting these videos together in an entertaining, informative and easy to follow way.

Liked and Subbed!

P.S. If you ever wanted to do a video on credit approval / risk assessment, you'd be my hero even MORE

[mathematicalninja]

I'm looking to solve problems in education sector using AI and your channel has helped me a lot in going forward with that idea. Thanks!

[MohitPratapSingh]

u r really killing it in this new series!! No content like this on the entire web.. let lone youtube !!

[woolfel]

the slide at minute 7:41, inference rule engines are already used for regulatory compliance like 1940Act and FSA. You can't really use DNN for compliance for a couple of reasons. The first is that lawyers have to validate the compliance system satisfies the regulation, second is that using an inference rule engine like JESS, CLIPS and JRules is several orders of magnitude more efficient than using DNN. The third reason is that sequential evaluation of trades doesn't work well for diversification rules. This is typically done during portfolio rebalance or pre-trade compliance check. I designed and built a compliance engine over a decade. For fraud detection, DNN has potential. I know there are firms actively working with DNN to build more dynamic fraud systems that needs less human intervention.

[muhammadsami]

Thanks, man. Right in time. I have said this in your previous videos and am saying it again. You are a great person with a big vision. We have so much in common. I am starting two A.I-based startups too (one after the other that have been made possible by 5 months of planning) and I would love that we work on it together.

[gustavomartinez]

A video that explains the current methods to understand why reinforcement learning or deep learning has taken that decision, ( causality) would be great!!!!. You are a very inspiring person, maybe we have a lot of differences, but you inspire a lot of people, so that's great!!!

[frankhaugen]

I work with ML in finance, automating invoice registration, (I don't do the AI, but I do supporting microservices). And yes, finance is where it's at. Good vid

[tiscottt]

very good presentation! silky explanations with technical background. Awesome job dude, Respect.

[bonifaceyogendran]

Great job man. I am already working on the AI model for FX market. You mentioned lot of things that I was not aware of.

[zionformulasmagicas]

Siraj, the format used in line 10 is deprecated:

from sklearn import preprocessing, cross_validation, svm

The correct format is:
from sklearn import preprocessing, svm
from sklearn.model_selection import cross_val_score as cross_validation

Obs.: I'm rapper (retired), and i love your songs.

[gerardomoscatelli]

Linear regression of a price as a function o time, everything goes up forever until it doesn't :)

[saitaro]

Siraj's forelock changed it's color. It's a sign. I close out.

[bryancopeland]

Been loving the the AI in Finance videos! Its sad that they dont teach you this in business school, at least not yet. Thank your Siraj you the man!

[oberyingaming]

Your videos are GOLD, Siraj ! Keep up the good work :)

[KadirErturk]

I think the biggest issue is that the data is not time invariant. Therefore none of the solution can work forever.
Plus, with Reinforcement learning the action should effect result, which is not the case in stock.

[houdamouttalib]

I already was needing that, thanx Siraj

[jgo]

Talk about timing! was looking for something like this for a PoC