Machine learning

Fintech are increasingly using machine learning which is a kind of artificial intelligence where computers can learn without being programmed, i.e. it performs better with experience, like how an ideal human being is expected to improve with experience. It is used in varied application like predictive analysis for credit defaults, fraud detection, algorithmic trading and personalised advice and interaction
Machine learning is a kind of artificial intelligence where computers have the ability to learn without being programmed, i.e. it performs better with experience, similar to how an ideal human being is expected to improve with experience
Much of machine learning overlaps with similar fields like data mining (where patterns are discovered rather than explicitly searched for) but focuses on prediction, based on previous data
Machine learning has found various applications in fintech including:
- Predictive Analysis for Credit defaults
- Improved decision making
- Fraud Detection and Identity Management
- Extraction of information from multiple sources
- Algorithmic trading
- Personalised advice and interaction
Some solutions
Predictive Analysis for Credit defaults
Prevention of loss is a key aspect of lending. Some of the previous models in predicting loss have been based on historical financial data when it is available. Fintech companies use larger data sets from different sources and analyse these to create prediction models that inevitably improve current credit risk models.
- Early warning credit card data breach and fraud reduction solution by focusing on merchants big data
- Diagnostic tool for company credit risk assessment and predicts default
Other companies who use this technology include LendingClub, Kabbage and Lendup
Improved decision making
Machine learning allows computers to process data far faster and more efficiently to improve decision making
- Early discovery of trending news that provides a time advantage for investment decisions.
- Using artificial intelligence systems to simplify and manage cash cycle management
- Statistically guide the decision making process of a customer to the most suitable product
Other companies that use this technology are Affirm, ZestFinance and BillGuard
Fraud Detection and Identity Management
Machine learning can assist by analysing patterns in historical transaction data and build a model that detect fraudulent patterns
- Algorithmic detection of credit card fraud for online transactions
- Real time fraud prevention through machine learning models analysing big data
Other companies that use this technology are Bionym, EyeVerify and BioCatch.
Extraction of information from multiple sources
Multiple independent data sets now exist, much of it user generated, e.g. social media. Several machine learning applications extract information from these sources and package them with insights that enhance primary client financial data
- Analyses social media of clients to identify triggers relating to banking
- Extract stock related conversations from social media and rates financial bloggers
- Sentiment analysis of news sources to enhance trading algorithms
Others companies that use this technology are Dataminr and AlphaSense
- Contact Terence Singh for specifics about any of the solutions mentioned above
- Challenges and Scans – get a 10% discount when you use the code “TrendReports” (http://campaigns.matchi.biz/discount/)
- Follow Matchi on Twitter and LinkedIn
More on Matchi: Matchi is a global fintech innovation match-making firm, since 2013. Matchi has worked with over 100 leading banks and insurance companies (FIs) around the world, and has a database of over 2,500 fintech firms. Matchi provides both a highly curated portal of fintech solutions, as well as bespoke projects for FI clients to source targeted fintech solutions aimed at the FIs focus areas / pain points. The global Matchi team has run Innovation Challenges and Market Scans for multiple FIs around the world, in markets as diverse as Canada, Japan and India, as well as searches in other geographies.