How to apply AI to small data problems - TechCrunch

How to apply AI to small data problems TechCrunch

Over the past decade or so, the digital version has a rest of data. This is inspiring for a number of reasons, but mainly in terms of how AI will be able to revitalize the enterprise.

However, in the world of B2B - the industry in which I am heavily involved - we still suffer from a lack of data, largely because the number of transactions is much lower compared to B2C. Therefore, in order for AI to fulfill its promise of enterprise innovation, it must be able to solve these small data problems as well. Fortunately, it can.

The problem is that many data scientists are turning to bad habits, creating self-fulfilling prophecies, which reduce the effectiveness of AI in small data situations - and ultimately adding hindering the influence of AI in advancing the enterprise.

The trick to properly applying AI to small data problems is in following proper data science practices and avoiding bad ones.

The term "self-fulfilling prophecy" is used in psychology, investment and elsewhere, but in the world of data science, it can simply be described as "predicting what is obvious." We see this when companies find a model that predicts what already works for them, sometimes even “by design,” and applies it to different situations.

For example, a sales company confirms that people who fill out their card online are more likely to buy than people who did not, so they market heavily to that group. They are predicting what is obvious!

READ  AWS can't rest - TechCrunch

Instead, they should implement models that help to do what is possible No works well - switching first time buyers who do not already have items in their cart. By opting for the latter - or predicting the unpredictable - this sales company will be far more likely to influence sales and gain new customers rather than just hold on. the same ones.

To trap creating self-fulfilling prophecies, here's the process you should follow to apply AI to small data problems:

  1. Enrich Your Data: When you find that you do not have a ton of custom data to work with, the first step is to enrich the data that you already have. This can be done by using external data to apply similar modeling. We’ll see this more than ever thanks to the rise of recommendation systems used by Amazon, Netflix, Spotify and more. Even if you only have one or two purchases on Amazon, they have so much information about products in the world and the people who buy them, that they can make a very accurate prediction of your next purchase. . If you are a B2B company that uses "one sided" to classify your contracts (eg, "large companies"), follow the example of Pandora and eliminate all customers with the most accurate ratings (e.g., song title, artist, sex singer, tune pick, beat, etc). The more you know about your data, the richer it will get. You can go from low data with low predictions to high-end experience with powerful predictions and recommendation models.

Related Posts

Deja una respuesta

Tu dirección de correo electrónico no será publicada.

Subir

We use cookies to ensure that we give the best user experience on our website. If you continue to use this site we will assume that you agree. More information