Scribd
Upload
66 views

Applications of AI and Data Mining in Trading

This document discusses applications of AI and data mining in trading. It describes how data mining can discover meaningful patterns in large datasets, including traditional datasets like historical prices and fundamentals as well as alternative datasets from sources like social media, business transactions, and sensor data. It also discusses how natural language processing and sentiment analysis can extract trading signals from text, audio, and social media. Finally, it provides examples of supervised learning techniques like decision trees and support vector machines as well as unsupervised learning and genetic algorithms that are used by firms to identify patterns and make trading decisions.

Uploaded by

ab
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
66 views

Applications of AI and Data Mining in Trading

This document discusses applications of AI and data mining in trading. It describes how data mining can discover meaningful patterns in large datasets, including traditional datasets like historical prices and fundamentals as well as alternative datasets from sources like social media, business transactions, and sensor data. It also discusses how natural language processing and sentiment analysis can extract trading signals from text, audio, and social media. Finally, it provides examples of supervised learning techniques like decision trees and support vector machines as well as unsupervised learning and genetic algorithms that are used by firms to identify patterns and make trading decisions.

Uploaded by

ab
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 8

Applications of AI and Data

Mining in Trading
Data Mining
• Discovering meaningful patterns, trends and correlations in large
datasets
• Big data – volume velocity and variety
• Big Data is often multi-dimensional and requires visualization tools
such as Tableau

Traditional datasets Alternative datasets

Company
Historical Prices Ratios Individual data Business data Sensor data
Fundamentals

Transactions ,
social media, reviews satellite images, CCTVs,
corporate data,
and news, web search traffic cams,
governmental agency
trends geolocation
data
Natural Language Processing
& Sentiment Analysis
• Computer analyses text and audio to extract
valuable trading signals
• Decipher sentiments like positive/negative
reactions towards a stock, product or
company
• Advanced computer power and programming
algorithms are allowing traders to “quantify
the unquantifiable”
• iSentium:
– Uses NLP to analyze unstructured social media
content and create Daily Directional Indicators on
stocks and ETFs
• Accern:
– Derives trading analytics such as sentiment, impact
and reliability from news stories, blogs, SEC filings and
social media
• Descartes Labs:
– Produces geospatial analytics that are used to track
firms in agriculture, logistics and commodities trade.
Supervised Learning
• JP Morgan’s Quantitative Research Team uses Random
Trees
– Decision trees find the optimal rule to forecast an
outcome based on a sequence of simple decision steps.
– Random Forests average simple decision tree models.
This yields better and more reliable forecasts
Supervised Learning
• Societe Generale
Electronic Trading uses
Support Vector
Machines
• This method maps
observations into a
multi-dimensional space
and separated into
categories using a
maximum margin
classifier
Unsupervised Learning
• Principle component analysis
– Find the main drivers from many independent
variables
• K-means
– Classify data into subgroups
– Data in the same group are similar.
– Develop different models for different groups
Genetic Algorithms
• Evolutionary biology -
survival of the fittest
• Failing techniques are Initialization of zero
generation individuals
disabled while successful
ones are retained, adapted
and evolved to fit new
market behavior
Selection of individuals
• Aidyia , a Hong Kong based Testing of efficiency
and profitability
with greater predictive
ability
algorithmic trading firm
uses Genetic Algorithms to
identify patterns and price
movements 2nd generation created

• Sentient Investment through Crossings,


Mutations and
Reproductions
Management uses
evolutionary intelligence to
make trading decisions

You might also like