3. Introduction
Name – Sanjay Kumar Kantilal Mota
Email – skmota@gmail.com
Qualification – Bachelor of commerce, MBA (Finance), PG Diploma in Computers
Certifications – PRINCE2 Practitioner, ITIL Foundation, Microsoft Certified Professional, Siebel
Certified Consultant
Work Experience – 26 years of IT experience working with Mastek, Deloitte, GENPACT, Wipro,
Satyam, Accenture
Geographic Experience – India, UK, USA, Japan
Industry Experience – 6 years of Capital Market related projects
4. Artificial Intelligence
SourceData
• Historical
Price Data
• EOD
• Intraday
DataMining
• Encoding
• Decoding
MachineLearning
Algorithms
• Training
• Testing
• Classification
• Regression
• Prediction
Trading
• Automated
or Manual
• Intra-day
• Positional
High Level Process Flow
5. Daily Predictions
Based on historical End of Day price data using Data Mining, Machine Learning and AI
14 years of Training Data and 6 years of Test Data used
Predicts if next day closing will be higher or lower than the next day
opening levels
Accuracy of 59% for Nifty, 9% more than by random guessing
Profit from 59% Accurate Prediction 39862.15 points (871 days)
Loss from 41% In-Accurate prediction 23963.5 points (604 days)
Net profit 15898.65 points (Total 1475 days)
For large volume Accuracy can be improved to 63% i.e. 13% more than
by random guessing
Can be used for Manual or Automated Intraday Trading
Prediction available after closing of the current day market and before
opening of the next day market
6. For Prediction of Closing > Opening
Buy at and below opening Level
Can also be used to trade in Binary Options
Spread out the trade by creating additional buy
trade when price falls by certain units e.g. every 0.05
points
For Prediction of Closing < Opening
Sell at and above opening level
Can also be used in Binary Options
Spread out the trade by creating
additional sell trade when price rises
by certain units e.g. every 0.05 points
Suggested Automated Trading Approach
Close all open positions before close of Trading
Incremental spread of Trades by certain units will ensure higher profitability on
profitable days (59%) and lesser loss on loss making days (41%)
Intraday trading will provide for higher leverage and no overnight carry over
7. -
Target Customers
Plan to work with very few Large Trading Companies
• Instruments with Large Market Capitalization and Liquidity
• Different World Markets (FTSE, DJIA, S&P500, NIFTY …)
• Large Trading Position
Profitability
• Plan to charge % of profit share
8. The Service is offered as it is on profit sharing basis without taking any responsivity for any trading risk
and trading loss.
Data for the instruments to be traded has to be provided by Customer
The process used for predictive analytics using Data Mining, Machine Learning and AI are my IP rights
and should be respected accordingly
I am free to deal with other companies to use predictive analytics using Data Mining, Machine
Learning and AI for other instruments in same or other markets
Accuracy varies for each instrument and needs to checked based on historical price data
Future prediction might or might not follow the past performance
Important Notes
9. Cracking
Wall Street
Can
technology
build a
better
Buffett?
List of Funds
or Trading
Firms Using
Artificial
Intelligence
How
Computers
Trawl a Sea
of Data for
Stock Picks -
WSJ
Battle of the
Quants
Quants turn
to AI for
market
insights
Investor
rush to
artificial
intelligence
is real deal -
FT.com
AI That
Picks Stocks
Better Than
the Pros |
MIT
Technology
Israel's
Financial
Algorithms
Gets $4B
Buy Bid,
Calcalist
Says
The Rise of
the Tech
Model May
Soon Make
You
Obsolete
BlackRock
and Google
in talks over
joint
venture -
FT.com
Big banks
invest
$13.5M in
machine
learning
startup ..
Goldman
Sachs Leads
$15 Million
Investment
in Kensho ..
Artificial
intelligence
is the next
big thing for
hedge funds
Introducing
Binatix, a
Deep-
Learning
Trading Firm
That's ...
Algorithmic
Trading:
Swing
Trading
Based on
Machine ..
Forex
Scandal
Drives Shift
to Algo
Trading -
Wall Street
...
Interesting Readings