The Fifth Elephant 2018

The seventh edition of India's best data conference

Deep portfolio: using neural networks for portfolio construction

Submitted by Anant Gupta (@anantguptadbl) on Tuesday, 27 March 2018

videocam_off

Technical level

Intermediate

Section

Full talk

Status

Confirmed & Scheduled

View proposal in schedule

Vote on this proposal

Login to vote

Total votes:  +14

Abstract

Deep Learning is a good concept and it is slowly transforming the face of data analysis. The world of finance has not been impervious to its reach. Although finance has its own models which have in place for decades ( Black Sholes, CAPM ) new methodologies are coming up to leverage the power of AI

For the past few decades Portfolio building has been guided by analytical methodologies. There have been several methodologies like CAPM, Apt, Factor Data etc. However, very recently there have been some papers which leverage Deep Learning for portfolio building. Through this Full Talk I will be presenting a deep learning application in the field of portfolio composition

Outline

There are many hidden factors that relate the universe of financial products which cannot be unearthed using APT. However, using Deep Neural Networks, we can hunt for the latent linkages between the financial products and use that for building your portfolio
The process can be divided into the following 3 tasks
1) Auto Encoding
This step will be used for creating a condensed map of the entire universe of financial products

2) Calibrating
This allows us to choose a particular target ( benchmark or a manual set of returns ) that we have in mind that we would like to create a portfolio for

3) Validation and Verification
This allows us to choose the appropriate condensed map of the universe of produces which will give the best calibration

There is a paper that talks about the potential applications of Deep

As part of the talk, I will be choosing the following examples
1) Benchmark the NIFTY Index
2) Benchmark a list of user defined returns

During the course of the talk, I will be using the following technologies and data
1) Python : Tensorflow, numpy
2) Jupyter : For coding/visualization
3) Datasets : Open Financial data from Quandl/Kaggle etc

Requirements

An interest in Finance and Statistics. If you are always looking out for new ways of investment then this will definitely be worthwhile

I will be doing the coding in Python

Speaker bio

Anant is part of Morgan Stanley for the past 7 years. He did his B.Tech ( Electrical ) from BITS Pilani

Slides

https://drive.google.com/open?id=1zvlEF1Tofox7UQAHN4tG2s55I84T5Hh_

Comments

  • 3
    Runjhun Kejriwal (@runsi) 7 months ago

    good job man, i will be looking forward to it

  • 2
    Sandip Gupta (@montylicious) 8 months ago

    Great going. Put the metal on the pedal. Yaalgar Ho!

  • 2
    Ashish Kumar Singh 8 months ago

    Awesome!! Anant

  • 2
    Neeta Pande (@neetapande) 8 months ago

    Great application of deep learning in Fintech!!

  • 2
    Umang Agrawal (@umang10) 7 months ago

    Great…gud going..

  • 1
    Zainab Bawa (@zainabbawa) Reviewer 7 months ago

    Anant, we need slides and preview video to evaluate this proposal. No proposals will be evaluated without both these details.

  • 1
    Anant Gupta (@anantguptadbl) Proposer 7 months ago

    Sure Zainab, I will get back to you with the deck by this weekend

  • 1
    Hari C M (@haricm) Reviewer 7 months ago

    Anant, any update on slides?

Login with Twitter or Google to leave a comment