Wednesday, May 25, 2016
Applications of Machine Learning Algorithms in Banking - entering the next era of Banking
Artificial Intelligence
,
Banking
,
Big Data
,
classification
,
clustering
,
DBS digibank
,
digibank
,
Hadoop
,
Machine Learning
,
recommendation
No comments
:
"People read around 10MB worth of material, hear 400MB and see 1MB of information every second per day" - Economist
In the world of excess information, it is becoming more and more difficult to look for things that we like to read, buy and do. We have always relied on opinions from other people to understand the nature of the things that we would like to have. Some of us would have watched movies based on the recommendation from our friends and cousins. Now, the world has so much to offer to everyone, according to their taste and this has made it difficult for people to always look for the opinion of others as it is difficult to find a person similar to your taste within your circle.
Image courtesy: Pixabay
Good news is that the world has gone to the next stage without explicitly telling us how they have replaced this friendly opinions and simply replaced it with a machine-based algorithms. For example, Netflix; when you have bought the subscription, you would have seen "recommendation for you" and you would have also noticed "because you have watched" section. Ever wondered how Netflix knew that you may want to watch next ? What's happening at the backend to provide you all these suggestions? The answer is "Machine Learning".
Machine learning is a subfield of computer science that evolved from the study of pattern recognition and computational learning theory in artificial intelligence. In 1959, Arthur Samuel defined machine learning as a "Field of study that gives computers the ability to learn without being explicitly programmed". ~ From Wikipedia
Machine learning belongs to Artificial intelligence on a higher level and Artificial Intelligence is achieved through "Machine Learning" and "Predictive Analysis". There are 3 major machine learning techniques to achieve the "Intelligent" state:
- Recommendation
- Classification
- Clustering
Netflix - Uses recommendation algorithm based on the personal preferences and the community preferences. Netflix keeps track of the movies watched by individuals and reads the pattern of what is watched next and suggests this to the new users.
Google Email - Uses classification algorithm to classify the spam and good emails. Once a substantial number of individuals mark an email as "Spam", the machine takes note of it and classify such email as SPAM for you and other users.
Google News - Uses clustering algorithm to display news from various sources based on the preferences of the clustered groups.
Machine learning techniques are used across all industries and Banking sector has entered the race now. The banks are now employing the data mining tools to understand the behaviour of the customer and recommend products based on the spending behaviour. Recently, the DBS bank has launched the "digibank" and DBS CEO Piyush Gupta, defined it as the "WhatsApp moment in banking". This digibank application is supported in the backend with the machine learning algorithm and it has a virtual assitant that will understand natural language such as "What is my account balance?".
"A bank is as successful as a community or a society is. So, it is important for the bank to think what is needed for the society".
To construct a successful bank or to retain the success or to helm, the banks have to come up with innovative products than the traditional loans and credit card offerings. In the modern world, success comes through innovation and timely understanding of the pressing needs of the society. Machine learning helps to understand the needs of an individual, group, community or a society. Thus, machine learning will help banks to offer innovative products specific to different customer segments.
Applications of Machine Learning Techniques in Banking, but not limited to:
- Recommendation based on the pattern of spending on their credit card / ATM cards
- Product suggestions based on the customer's behaviour in a given product catalogue / website
- Answering customers based on their natural language queries. Example: "what is my credit card due date?"
- Predictive analytics to find what would be the customers' next steps or actions if he has purchased a product or he has booked a flight
- Use of clustering algorithm to group bank customers based on their spending nature and offer products specific to that group/cluster.
- Imagine a virtual reception to receive the customer and assist who is willing to invest, but all just on his mobile phone
- Predictive analysis of customers missing something (example, a flight) and what would be their immediate need
- Suggest customers of a particular third party product, restaurant, bar that is affiliated with the bank and offers great discounts. Note that, the bank grows as its affiliations grow.
- Advise customers if their card is swiped at two different geo locations using a push notification and suggest him next steps
- Analysis of delinquent customers and take necessary actions
- Understand the customers' bill payment cycles for all their subscriptions and advice on a timely fashion using a push notification.
- Use of clustering algorithm to group bank customers and suggest them to partner among themselves for their betterment and thus bringing growth to the bank.
With machine learning and its implementation, Banking is not same as it was earlier and it has changed and it is evolving to use all better practices from the technology industry. Machine learning is not something new, but Banks have started reaping the potentials of it. Earlier, you relied on a Banks' representative to help you choose the product or services, but not anymore as a simple digibank application may help you find what you need.
Trivial:
While Machine Learning is an exciting thing around with greater potentials, machine learning requires a great source of data and it needs large storage capacity to process the data that will be used for data mining and predictive analysis. Thus, Big Data or Hadoop is an essential platform to build your machine learning applications or solutions.
While Machine Learning is an exciting thing around with greater potentials, machine learning requires a great source of data and it needs large storage capacity to process the data that will be used for data mining and predictive analysis. Thus, Big Data or Hadoop is an essential platform to build your machine learning applications or solutions.
Reference/Source:
Subscribe to:
Post Comments
(
Atom
)
No comments :
Post a Comment