Cookies help us display personalized product recommendations and ensure you have great shopping experience.

By using this site, you agree to the Privacy Policy and Terms of Use.
Accept
SmartData CollectiveSmartData Collective
  • Analytics
    AnalyticsShow More
    sales and data analytics
    How Data Analytics Improves Lead Management and Sales Results
    9 Min Read
    data analytics and truck accident claims
    How Data Analytics Reduces Truck Accidents and Speeds Up Claims
    7 Min Read
    predictive analytics for interior designers
    Interior Designers Boost Profits with Predictive Analytics
    8 Min Read
    image fx (67)
    Improving LinkedIn Ad Strategies with Data Analytics
    9 Min Read
    big data and remote work
    Data Helps Speech-Language Pathologists Deliver Better Results
    6 Min Read
  • Big Data
  • BI
  • Exclusive
  • IT
  • Marketing
  • Software
Search
© 2008-25 SmartData Collective. All Rights Reserved.
Reading: When sharing isn’t a good idea
Share
Notification
Font ResizerAa
SmartData CollectiveSmartData Collective
Font ResizerAa
Search
  • About
  • Help
  • Privacy
Follow US
© 2008-23 SmartData Collective. All Rights Reserved.
SmartData Collective > Big Data > Data Mining > When sharing isn’t a good idea
Data Mining

When sharing isn’t a good idea

TimManns
TimManns
6 Min Read
SHARE

Ensemble models seem to be all the buzz at the moment. The NetFlix prize was won by a conglomerate of various models and approaches that each excelled in subsets of the data.

A number of data miners have presented findings based upon using simple ensembles that use the mean prediction of a number of models. I was surprised that some form of weighting isn’t commonly used, and that a simple mean average of multiple models could yield such an improvement in the global predictive power. It kinda reminds me of Gestalt theory phrase “The whole is greater than the sum of the parts.” It’s got me thinking, when it is best not to share predictive power. What if one model is the best? There is also a ton of considerations regarding scalability and trade-off between additional processing, added business value, and practicality (don’t mention random forests to me…), but we’re pretend those don’t exist for the purpose of this discussion 🙂

So this has got me thinking do ensembles work best in situations where there are clearly different sub-populations of customers. For example, Netflix is in the retail space, with many customers that rent the same popular blockbuster movies, and a moderate .. …



Ensemble models seem to be all the buzz at the moment. The NetFlix prize was won by a conglomerate of various models and approaches that each excelled in subsets of the data.

More Read

“While touring IBM’s Innovation lab at Lotusphere last…
Open standards for data mining and the need for training material
Interview –Michael Zeller CEO,Zementis
How to compare two audio files quality wise?
Top 10 analytic mistakes

A number of data miners have presented findings based upon using simple ensembles that use the mean prediction of a number of models. I was surprised that some form of weighting isn’t commonly used, and that a simple mean average of multiple models could yield such an improvement in the global predictive power. It kinda reminds me of Gestalt theory phrase “The whole is greater than the sum of the parts.” It’s got me thinking, when it is best not to share predictive power. What if one model is the best? There is also a ton of considerations regarding scalability and trade-off between additional processing, added business value, and practicality (don’t mention random forests to me…), but we’re pretend those don’t exist for the purpose of this discussion 🙂

So this has got me thinking do ensembles work best in situations where there are clearly different sub-populations of customers. For example, Netflix is in the retail space, with many customers that rent the same popular blockbuster movies, and a moderate number of customers that rent rarer (or far more diverse, i.e., long tail) movies. I haven’t looked at the Netflix data, so I’m guessing that most customers don’t have hundreds of transactions, so generalising the correct behaviour of the masses to specific customers is important. Netflix data on any specific customer could be quite scant (in terms of rents/transactions). In other industries such as telecom, there are parallels; customers can also be differentiated by nature of communication (voice calls, sms calls, data consumption etc) just like types of movies. Telecom is mostly about quantity though (customer x used to make a lot of calls etc). More importantly there is a huge amount of data about each customer, often with many hundreds of transactions per customer. There is therefore relatively lesser reliance upon supporting behaviour of the masses (although it helps a lot) to understand any specific customer.

Following this logic, I’m thinking that ensembles are great at reducing the error of incorrectly applying insights derived from the generalised masses to those weirdos that rent obscure sci-fi movies! Combining models that explain sub-populations very well makes sense, but what if you don’t have many sub-populations (or can identify and model their behaviour with one model).

But you may shout “hey, what about the KDD Cup.” Yes, the recent KDD Cup challenge (anonymous featureless telecom data from Orange) was also a won by an ensemble of over a thousand models created by IBM Research. I’d like to have had some information about what the hundreds of columns respresented, and this might have helped better understand the Orange data and build more insightful and performing models. Aren’t ensemble models used in this way simply a brute force approach to over learn the data? I’d also really like to know how the performance of the winning entry tracks over the subsequent months for Orange.

Well, I haven’t had a lot of success in using ensemble models in the telecom data I work with, and I’m hoping it is more a reflection of the data than any ineptitude on my part. I’ve tried simply building multiple models on the entire dataset and averaging the scores, but this doesn’t generate much additional improvement (granted on already good models, and I already combine K-means and Neural Nets on the whole base). During my free time I’m just starting to try splitting the entire customer base into dozens of small sub-populations and building a Neural Net model on each, then combining the results and seeing if that yields an improvement. It’ll take a while.

Thoughts?

Link to original post

Share This Article
Facebook Pinterest LinkedIn
Share

Follow us on Facebook

Latest News

sales and data analytics
How Data Analytics Improves Lead Management and Sales Results
Analytics Big Data Exclusive
ai in marketing
How AI and Smart Platforms Improve Email Marketing
Artificial Intelligence Exclusive Marketing
AI Document Verification for Legal Firms: Importance & Top Tools
AI Document Verification for Legal Firms: Importance & Top Tools
Artificial Intelligence Exclusive
AI supply chain
AI Tools Are Strengthening Global Supply Chains
Artificial Intelligence Exclusive

Stay Connected

1.2kFollowersLike
33.7kFollowersFollow
222FollowersPin

You Might also Like

What is Hadoop?

5 Min Read

Signals sinchronizer and its role in automated perceptural voice quality testing

1 Min Read

Analytics BS: 3 Questions to Spot It

11 Min Read

Why Google Chrome has a bad market share ?

1 Min Read

SmartData Collective is one of the largest & trusted community covering technical content about Big Data, BI, Cloud, Analytics, Artificial Intelligence, IoT & more.

AI chatbots
AI Chatbots Can Help Retailers Convert Live Broadcast Viewers into Sales!
Chatbots
ai chatbot
The Art of Conversation: Enhancing Chatbots with Advanced AI Prompts
Chatbots

Quick Link

  • About
  • Contact
  • Privacy
Follow US
© 2008-25 SmartData Collective. All Rights Reserved.
Go to mobile version
Welcome Back!

Sign in to your account

Username or Email Address
Password

Lost your password?