Skip to main content

Correlated Cross-Occurrence (CCO): How to make data behave



Cross-occurrence allows us to ask the question: are 2 events correlated.

 To use the Ecom example, purchase is the conversion or primary action, a detail page view might be related but we must test each cross-occurrence to make sure. I know for a fact that with many ecom datasets it is impossible to treat these events as the same thing and get anything but a drop in quality of recommendations (I’ve tested this). People that use the ALS recommender in Spark’s MLlib sometimes tell you to weight the view less than the purchase. But this is nonsense (again I’ve tested this). What is true is that *some* views lead to purchases and others do not. So treating them all with the same weight is pure garbage.

 What CCO does is find the views that seem to lead to purchase. It can also find category-preferences that lead to certain purchases, as well as location-preference (triggered by a purchase when logged in from some location).  And so on. Just about anything you know about users or can phrase as a possible indicator of user taste can be used to get lift in quality of recommendation.

 So in the example below purchase history is the conversion action, likes, and downloads are secondary actions looked at as cross-occurrences. Note that we don’t need to have the same IDs for all actions. 

 BTW to illustrate how powerful this idea is, I have a client that sells one item a year on average to a customer. It’s a very big item and has a lifetime of one year. So using ALS you could only train on the purchase and if you were gathering a year of data there would be precious little training data. Also when you have a user with no purchase it is impossible to recommend. ALS fails on all users with no purchase history. However with CCO, all the user journey and any data about the user you can gather along the way can be used to recommend something to purchase. So this client would be able to recommend to only 20% of their returning shoppers with ALS and those recs would be low of quality based on only one event far in the past. CCO using all the clickstream (or important parts of it) can do quite well.

 This may seem an edge case but only in degree, every ecom app has data they are throwing away and CCO addresses this.
A way to compare 2 events at the individual level. The comparison is called the Log-Likelihood Ratio or LLR. This would allow us to look across all page views and see which correlated with which purchases. , not all page view are created equal but we now had a way to find the important ones!

Comments

Post a Comment

Popular posts from this blog

Real-Time Talk: Windows 10 IoT Core Background Tasks and ASP.NET Core Web Apps

Display useful information from your Windows 10 IoT Core application in an ASP.NET Core web app, essential for integrating IoT data into a solution. Windows 10 IoT background task talk with a web application using WebSockets. Problems As my path to this solution has been troublesome, I am listing here the main problems I faced so my dear readers have a better idea of dead-end streets along the way: I was not able to make the ASP.NET Core web application run under a Windows 10 IoT background service. I found no information about when or if it will be supported in the near future. ASP.NET MVC and ASP.NET Core have different SignalR implementations. I was not able to make a SignalR client for .NET Core work with SignalR hosted on a web application. I was able to make things work by directly using a WebSocket. It’s not as nice a solution as I had in my mind, but it works until things get better. Making the Background Task and Web Application Talk I worked out sim...
Post a Comment
Read more

Python and Parquet Performance

In Pandas, PyArrow, fastparquet, AWS Data Wrangler, PySpark and Dask. This post outlines how to use all common Python libraries to read and write Parquet format while taking advantage of  columnar storage ,  columnar compression  and  data partitioning . Used together, these three optimizations can dramatically accelerate I/O for your Python applications compared to CSV, JSON, HDF or other row-based formats. Parquet makes applications possible that are simply impossible using a text format like JSON or CSV. Introduction I have recently gotten more familiar with how to work with  Parquet  datasets across the six major tools used to read and write from Parquet in the Python ecosystem:  Pandas ,  PyArrow ,  fastparquet ,  AWS Data Wrangler ,  PySpark  and  Dask . My work of late in algorithmic trading involves switching between these tools a lot and as I said I often mix up the APIs. I use Pandas and PyArrow for in-RAM comput...
Post a Comment
Read more

DANGER (STILL) LURKS IN THE INTERNET OF THINGS (IOT)

All of the media, both “mainstream” and “tech”, has gushed over all of the new appliances and devices that are now in the category of what we would call the Internet of Things.  Items like home security,  home lighting, and refrigerators, to name a few. There are many advantages to having connected appliances and devices, Threats that can and will be exploited if unsuspecting users don’t secure them.   Two “layers” of security : The first layer of offering  is a security API that will provide [a way] to easily do a virtual patch, to prevent a remote attack, for example . . . the third layer is cloud: IoT cannot do anything without the cloud.  Most data is sent to the cloud and you will need to have proper protection and make sure the cloud is always available. In both situation Users are vulnerable, mostly due to their own apathy.  Users often either don’t know how to patch their own machines (and in this case, devices) or have glanced...
Post a Comment
Read more