Overview of the Supervised Machine Learning in Finance

Loading...

En provenance du cours de New York University Tandon School of Engineering

Guided Tour of Machine Learning in Finance

148 notes

New York University Tandon School of Engineering

Guided Tour of Machine Learning in Finance

148 notes

Cours 1 sur 4 dans Specialization Machine Learning and Reinforcement Learning in Finance

This course aims at providing an introductory and broad overview of the field of ML with the focus on applications on Finance. Supervised Machine Learning methods are used in the capstone project to predict bank closures. Simultaneously, while this course can be taken as a separate course, it serves as a preview of topics that are covered in more details in subsequent modules of the specialization Machine Learning and Reinforcement Learning in Finance. The goal of Guided Tour of Machine Learning in Finance is to get a sense of what Machine Learning is, what it is for and in how many different financial problems it can be applied to. The course is designed for three categories of students: Practitioners working at financial institutions such as banks, asset management firms or hedge funds Individuals interested in applications of ML for personal day trading Current full-time students pursuing a degree in Finance, Statistics, Computer Science, Mathematics, Physics, Engineering or other related disciplines who want to learn about practical applications of ML in Finance Experience with Python (including numpy, pandas, and IPython/Jupyter notebooks), linear algebra, basic probability theory and basic calculus is necessary to complete assignments in this course.

À partir de la leçon

Mathematical Foundations of Machine Learning

Generalization and a Bias-Variance Tradeoff6:32

The No Free Lunch Theorem7:46

Overfitting and Model Capacity8:20

Linear Regression7:57

Regularization, Validation Set, and Hyper-parameters10:50

Overview of the Supervised Machine Learning in Finance3:58

DataFlow and TensorFlow10:58

A First Demo of TensorFlow11:08

Linear Regression in TensorFlow10:32

Rencontrer les enseignants

Igor Halperin

An old saying goes, if you have a hammer, everything looks like a nail.

And that's exactly what you will get by the end of this week.

You will get that machine learning hammer.

The hammer is neural networks combined with TensorFlow, a powerful library for

Machine Learning open sourced by Google in November 2015.

Chances are that you might have heard about or used these tools before.

But if not, then I think that the best time and

place to learn about these tools will be now and here.

So here's what you will learn in this week.

Let's start with our first lesson of this week.

And first, you will get familiar in this lesson with our main tool, TensorFlow.

We will start with a review of how TensorFlow implements what is called

a computational graph.

Then we will open our first TensorFlow Jupyter Notebook and

see how these things work by looking at very simple operations.

Next, you will see how simple linear regression can be implemented

TensorFlow and compare it with how it's done with other Python packages.

And after that, you will learn how neural nets are organized and

how linear regression is nothing but a very special and simple neural network.

You will also see why neural networks can

be seen as a kind of machine learning hammer.

After you learn about neural networks, you will learn about methods used for

their training.

And here, you will meet some of main power courses of modern machine learning,

namely, the Gradient Descent method and its relatives,

such as Stochastic Gradient Descent.

We will see they're working both in theory and in practice, that is,

in slides and in Jupyter Notebooks together.

Here, we'll see how it works first in a linear regression and

then neural network regression, implemented to TensorFlow.

All these will be done using some simple non-financial data, so

that we will not get lost litigating in several new topics simultaneously.

All these topics will be presented in the first lesson.

Then after we get such a powerful hammer, we will start to look for financial nails.

And the first nail you will see is a very classical financial problem,

namely the problem of predicting earnings per share, or EPS for short.

Once we have our tools such as TensorFlow and other packages,

such as scikit-learn and statsmodels, we will be able to try both linear and

nonlinear neural network regression on this problem.

And after that, we will turn to classification methods.

We will start with looking at how machine learning deals with probabilistic models.

And then we will talk about maximum likelihood estimation and

related methods such as Maximum A-Posteriori method.

We will also talk about the notion of relative entropy, also known as KL

divergence, which is very important for understanding for machine learning.

Then you'll see how all this machinery applies to classification problems.

You will learn about one of the most powerful approach to classification called

logistic regression.

You will also learn how it can be implemented in scikit-learn and

TensorFlow.

You will see how logistic regression is just another special

case of our general hammer called neural nets.

And finally, our hammer will find its second financial nail

that will be a problem of predicting bank failures.

As you will see, our hammer will be pretty good for this type of nails.

So this is what you will master this week.

Explorer notre catalogue

Rejoignez-nous gratuitement et obtenez des recommendations, des mises à jour et des offres personnalisées.

Coursera propose un accès universel à la meilleure formation au monde, en partenariat avec des universités et des organisations du plus haut niveau, pour proposer des cours en ligne.

© 2018 Coursera Inc. Tous droits réservés.

Télécharger dans l'App Store

Disponible sur Google Play