Markov process fits into many real life scenarios. Any sequence of event that can be approximated by Markov chain assumption, can be predicted using Markov chain algorithm. In the last article, we explained What is a Markov chain and how can we represent it graphically or using Matrices. In this article, we will go a step further and leverage this technique to draw useful business inferences.
[stextbox id=”section”] Business Case [/stextbox]
“Krazy Bank”, deals with both asset and liability products in retail bank industry. A big portfolio of the bank is based on loans. These loans make the majority of the total revenue earned by the bank. Hence, it is very essential for the bank to find the proportion of loans which have a high propensity to be paid in full and those which will finally become Bad loans. “Krazy Bank” has hired you as a consultant to come up with these scoping numbers.
All the loans, which have been issued by “Krazy Bank” can be classified into four categories :
[stextbox id=”section”] Short Note on Absorbing nodes [/stextbox]
Absorbing nodes in a Markov chain are the possible end states. All nodes in Markov chain have an array of transitional probability to all other nodes and themselves. But, absorbing nodes have no transitional probability to any other node. Hence, if any individual lands up to this state, he will stick to this node for ever. Let’s take a simple example. We are making a Markov chain for a bill which is being passed in parliament house. It has a sequence of steps to follow, but the end states are always either it becomes a law or it is scrapped. These two are said to be absorbing nodes. For the loans example, bad loans and paid up loans are end states and hence absorbing nodes.
[stextbox id=”section”] Transition diagram [/stextbox]
You have done a thorough research on past trends of loan cycle and based on past trends here is the Markov chain you observed (for a period of 1 year):
Explaining absorbing nodes becomes simple from this diagram. As you can see, Paid up and bad loans transition only to themselves. Hence, whatever path a process takes, if it lands up to one of these two states, it will stay there for ever. In other words, a bad loan cannot become paid up, risky or good loan ever after. Same is true with paid up loans.
[stextbox id=”section”] Transition calculations [/stextbox]
Once, we have 1 year transition probability, we can convert prediction algorithm to simple matrix multiplication. Currently the portfolio has 60% Good and 40% Risky loans. We look forward to calculate, how many of these loans will be finally paid up in full? Using 1 year transition probability, we can estimate the number of loans falling into each of the four bucket 1 year down the line.
Here are some interesting insights from this calculation. We can expect 15% of the loans to be paid up in this year and 16% being ending up as bad loans. As the % of bad loans seem to be on a higher side, it will be beneficial to identify these loans and make adequate interventions. Pin pointing to these 15% is not possible using simple Markov chain, but same is possible using a Latent Markov model.
Now, to make a prediction for 2 years, we can use the same transition matrix. This time the initial proportions will the final proportions of last calculation. Transition probability generally do not change much. This is because, it is based on several time points in past.
If we keep on repeating this exercise, we see the proportion matrix converges. Following is the converged matrix. Note that, multiplying it with transition matrix makes no change to proportions.
These are the proportion numbers we were looking for. 54% of the current loans will be paid up in full but 46% will default. Hence doing this simple exercise lead us to such an important conclusion that the current portfolio exposes bank to a very high risk.
[stextbox id=”section”] Loan transition deep dive [/stextbox]
We have already seen the stationary point proportions for the portfolio. Something which will be of interest to us next is that what proportion of Good loans land up being paid up in full. For this we can start with an initial proportion split of Good – 100% and rest – 0%. The final converged matrix is as follows:
Here are some interesting insights. If the entire portfolio was built of good loans, only 23% of loans would have defaulted against 46% for current portfolio. Hence, we will expect a very high proportion of risky loans will show default. We can find this using a simple transition calculation using Risky – 100% and rest – 0%. Following is the final converged matrix :
80% of such loans will default. Hence, our classification of Risky and Good separates out propensity to default pretty nicely.
[stextbox id=”section”] End Notes [/stextbox]
In this article, we saw how Markov chain can be used to find out multiple insights and make good predictions on an overall level. There are many other processes which can be explained using Markov chain. In such cases, Markov chain algorithm will give you number of insights and will serve as a very handy forecasting tool. However, Markov chain can only make forecast on segment level and not make prediction on customer level. In case you need to make customer level forecast, you need a Latent Markov model and not a simple Markov model.
Did you find the article useful? Are you aware of any other real life Markov process? Can Markov chain be used in that process to bring out interesting insights? Did this article solve any of your existing problems? Have you used simple Markov chain in any other context ?
Tavish Srivastava, co-founder and Chief Strategy Officer of Analytics Vidhya, is an IIT Madras graduate and a passionate data-science professional with 8+ years of diverse experience in markets including the US, India and Singapore, domains including Digital Acquisitions, Customer Servicing and Customer Management, and industry including Retail Banking, Credit Cards and Insurance. He is fascinated by the idea of artificial intelligence inspired by human intelligence and enjoys every discussion, theory or even movie related to this idea.
Introduction to Markov chain : simplified! (wit...
Markov chain : Mathematical formulation, Intuit...
A Comprehensive Guide on Markov Chain
How to interpret hidden state in Latent Markov ...
An Academic Overview of Markov Chain
Getting your clustering right (Part I)
Generating Fantasy Titles with a Markov Chain
Learn Analytics using a business case study : P...
A Beginner’s Guide to Channel Attribution...
Is survival analysis the right model for you?
We use cookies essential for this site to function well. Please click to help us improve its usefulness with additional cookies. Learn about our use of cookies in our Privacy Policy & Cookies Policy.
Show details
This site uses cookies to ensure that you get the best experience possible. To learn more about how we use cookies, please refer to our Privacy Policy & Cookies Policy.
It is needed for personalizing the website.
Expiry: Session
Type: HTTP
This cookie is used to prevent Cross-site request forgery (often abbreviated as CSRF) attacks of the website
Expiry: Session
Type: HTTPS
Preserves the login/logout state of users across the whole site.
Expiry: Session
Type: HTTPS
Preserves users' states across page requests.
Expiry: Session
Type: HTTPS
Google One-Tap login adds this g_state cookie to set the user status on how they interact with the One-Tap modal.
Expiry: 365 days
Type: HTTP
Used by Microsoft Clarity, to store and track visits across websites.
Expiry: 1 Year
Type: HTTP
Used by Microsoft Clarity, Persists the Clarity User ID and preferences, unique to that site, on the browser. This ensures that behavior in subsequent visits to the same site will be attributed to the same user ID.
Expiry: 1 Year
Type: HTTP
Used by Microsoft Clarity, Connects multiple page views by a user into a single Clarity session recording.
Expiry: 1 Day
Type: HTTP
Collects user data is specifically adapted to the user or device. The user can also be followed outside of the loaded website, creating a picture of the visitor's behavior.
Expiry: 2 Years
Type: HTTP
Use to measure the use of the website for internal analytics
Expiry: 1 Years
Type: HTTP
The cookie is set by embedded Microsoft Clarity scripts. The purpose of this cookie is for heatmap and session recording.
Expiry: 1 Year
Type: HTTP
Collected user data is specifically adapted to the user or device. The user can also be followed outside of the loaded website, creating a picture of the visitor's behavior.
Expiry: 2 Months
Type: HTTP
This cookie is installed by Google Analytics. The cookie is used to store information of how visitors use a website and helps in creating an analytics report of how the website is doing. The data collected includes the number of visitors, the source where they have come from, and the pages visited in an anonymous form.
Expiry: 399 Days
Type: HTTP
Used by Google Analytics, to store and count pageviews.
Expiry: 399 Days
Type: HTTP
Used by Google Analytics to collect data on the number of times a user has visited the website as well as dates for the first and most recent visit.
Expiry: 1 Day
Type: HTTP
Used to send data to Google Analytics about the visitor's device and behavior. Tracks the visitor across devices and marketing channels.
Expiry: Session
Type: PIXEL
cookies ensure that requests within a browsing session are made by the user, and not by other sites.
Expiry: 6 Months
Type: HTTP
use the cookie when customers want to make a referral from their gmail contacts; it helps auth the gmail account.
Expiry: 2 Years
Type: HTTP
This cookie is set by DoubleClick (which is owned by Google) to determine if the website visitor's browser supports cookies.
Expiry: 1 Year
Type: HTTP
this is used to send push notification using webengage.
Expiry: 1 Year
Type: HTTP
used by webenage to track auth of webenagage.
Expiry: Session
Type: HTTP
Linkedin sets this cookie to registers statistical data on users' behavior on the website for internal analytics.
Expiry: 1 Day
Type: HTTP
Use to maintain an anonymous user session by the server.
Expiry: 1 Year
Type: HTTP
Used as part of the LinkedIn Remember Me feature and is set when a user clicks Remember Me on the device to make it easier for him or her to sign in to that device.
Expiry: 1 Year
Type: HTTP
Used to store information about the time a sync with the lms_analytics cookie took place for users in the Designated Countries.
Expiry: 6 Months
Type: HTTP
Used to store information about the time a sync with the AnalyticsSyncHistory cookie took place for users in the Designated Countries.
Expiry: 6 Months
Type: HTTP
Cookie used for Sign-in with Linkedin and/or to allow for the Linkedin follow feature.
Expiry: 6 Months
Type: HTTP
allow for the Linkedin follow feature.
Expiry: 1 Year
Type: HTTP
often used to identify you, including your name, interests, and previous activity.
Expiry: 2 Months
Type: HTTP
Tracks the time that the previous page took to load
Expiry: Session
Type: HTTP
Used to remember a user's language setting to ensure LinkedIn.com displays in the language selected by the user in their settings
Expiry: Session
Type: HTTP
Tracks percent of page viewed
Expiry: Session
Type: HTTP
Indicates the start of a session for Adobe Experience Cloud
Expiry: Session
Type: HTTP
Provides page name value (URL) for use by Adobe Analytics
Expiry: Session
Type: HTTP
Used to retain and fetch time since last visit in Adobe Analytics
Expiry: 6 Months
Type: HTTP
Remembers a user's display preference/theme setting
Expiry: 6 Months
Type: HTTP
Remembers which users have updated their display / theme preferences
Expiry: 6 Months
Type: HTTP
Used by Google Adsense, to store and track conversions.
Expiry: 3 Months
Type: HTTP
Save certain preferences, for example the number of search results per page or activation of the SafeSearch Filter. Adjusts the ads that appear in Google Search.
Expiry: 2 Years
Type: HTTP
Save certain preferences, for example the number of search results per page or activation of the SafeSearch Filter. Adjusts the ads that appear in Google Search.
Expiry: 2 Years
Type: HTTP
Save certain preferences, for example the number of search results per page or activation of the SafeSearch Filter. Adjusts the ads that appear in Google Search.
Expiry: 2 Years
Type: HTTP
Save certain preferences, for example the number of search results per page or activation of the SafeSearch Filter. Adjusts the ads that appear in Google Search.
Expiry: 2 Years
Type: HTTP
Save certain preferences, for example the number of search results per page or activation of the SafeSearch Filter. Adjusts the ads that appear in Google Search.
Expiry: 2 Years
Type: HTTP
Save certain preferences, for example the number of search results per page or activation of the SafeSearch Filter. Adjusts the ads that appear in Google Search.
Expiry: 2 Years
Type: HTTP
These cookies are used for the purpose of targeted advertising.
Expiry: 6 Hours
Type: HTTP
These cookies are used for the purpose of targeted advertising.
Expiry: 1 Month
Type: HTTP
These cookies are used to gather website statistics, and track conversion rates.
Expiry: 1 Month
Type: HTTP
Aggregate analysis of website visitors
Expiry: 6 Months
Type: HTTP
This cookie is set by Facebook to deliver advertisements when they are on Facebook or a digital platform powered by Facebook advertising after visiting this website.
Expiry: 4 Months
Type: HTTP
Contains a unique browser and user ID, used for targeted advertising.
Expiry: 2 Months
Type: HTTP
Used by LinkedIn to track the use of embedded services.
Expiry: 1 Year
Type: HTTP
Used by LinkedIn for tracking the use of embedded services.
Expiry: 1 Day
Type: HTTP
Used by LinkedIn to track the use of embedded services.
Expiry: 6 Months
Type: HTTP
Use these cookies to assign a unique ID when users visit a website.
Expiry: 6 Months
Type: HTTP
These cookies are set by LinkedIn for advertising purposes, including: tracking visitors so that more relevant ads can be presented, allowing users to use the 'Apply with LinkedIn' or the 'Sign-in with LinkedIn' functions, collecting information about how visitors use the site, etc.
Expiry: 6 Months
Type: HTTP
Used to make a probabilistic match of a user's identity outside the Designated Countries
Expiry: 90 Days
Type: HTTP
Used to collect information for analytics purposes.
Expiry: 1 year
Type: HTTP
Used to store session ID for a users session to ensure that clicks from adverts on the Bing search engine are verified for reporting purposes and for personalisation
Expiry: 1 Day
Type: HTTP
Cookie declaration last updated on 24/03/2023 by Analytics Vidhya.
Cookies are small text files that can be used by websites to make a user's experience more efficient. The law states that we can store cookies on your device if they are strictly necessary for the operation of this site. For all other types of cookies, we need your permission. This site uses different types of cookies. Some cookies are placed by third-party services that appear on our pages. Learn more about who we are, how you can contact us, and how we process personal data in our Privacy Policy.
Edit
Resend OTP
Resend OTP in 45s
Hi Tavish, Thank you for sharing this. It's interesting and very useful. I have one clarification how do we get the "Final state transition matrix" 4x4 matrix?. Sorry if i am missing something here. Regards, S.S.Pradeep
Pradeep, The case discussed in this article has a constant transition probability matrix. Do you can use the same matrix throughout. Tavish
Hi Tavish, I have recently enrolled for a course in BA and trying to grasp as much information as possible. This Markov Chain problem correlates with some of the current issues in my Organization. Waiting for your future publications. Thanks & Regards, Koushik
Koushik, Can you illustrate the problem you are referring to?
Tavish this is interesting. can you elaborate on the deeep dive section.. how you arrived at the final converged matrix You mention "For this we can start with an initial proportion split of Good – 100% and rest – 0%. " So would the initial matrix be good risky bad paid up [100 0 0 0] If that is the initial matrix how do you arrive at good risky bad paid up [0 0 23 76] Thanks for your papers. they are very good for novices like me. Uday
Uday, The converged matrix is found by multiplying the transition matrix many times to the the current state. This is basically the final proportion of state distribution. You are right the initial matrix will be [100 0 0 0]. Keep on multiplying the transition matrix to this matrix and you will get [0 0 23 76]. Hope this helps. Tavish