The Easiest Explanation of Algorithms You Will Ever Read

We’ve all experienced the case whereby we are window-shopping online, or searching for something in Google, and next thing you know, that particular search is everywhere! Instagram, Facebook, Jumia, Konga, YouTube…

I’m sure you’ve had that creepy feeling like you’re being spied on, or tracked. Well, it’s actually all as a result of what is known as Recommendation Systems (the algorithm!).

Recommendation System (RS)? What’s that?

Recommendation systems are actually an application of artificial intelligence that use filtering practices to determine your preferences for content and products you actually care about.

RS decides what posts you see in your timeline, or the friends you should follow/add, or the videos you should watch on YouTube.

RS is the way services like Netflix, YouTube, Instagram, Spotify, and Jumia can be so spot on, or attuned with your wants and needs.

Ok, so how does it work?

There are basically 4 ways that recommendation systems work and these are: Content-Based Filtering, Popularity Filtering, Collaborative Filtering, and Hybrid Filtering. 

Content-Based Filtering

In this method, the engine tries to get initial information about the user when they make a new profile. In the process, the engine compares the items that the user has shown they like with the items the user didn’t like and then looks for similarities. Items similar to the liked ones will be recommended to the user.

Popularity Filtering

This is actually the easiest way to build an RS, by suggesting what is already popular. E.g. the items that are mostly sold in an online shop.

Collaborative Filtering

This method understands that two users have similar tastes if they have the same, or almost equivalent liked items in common. In this case things are suggested to the user based on what things people in their circle like.

Hybrid Filtering

This method for RS takes all the above methods and mixes it in different combinations. This is what most systems use today, because it doesn’t have the shortcomings the other methods have, such a ‘cold start’ (a shortcoming of content-based filtering in the case that the user is new and has no recorded likes or dislikes) or sparsity problem. 

Why do businesses and social media platforms invest so much in RS?

1. It generates traffic and engages more customers.
2. These engines automate processes, and thus reduce the workload of IT staff.
3. It brings about a personalised feel and customer satisfaction.
4. It provides information and reports that are relevant.

Just consider how you enjoy using YouTube or Netflix because it always recommends videos you enjoy, and how this encourages you to keep using it. Now you know why companies invest in RS algorithms.

Now wasn’t that a simple explanation?

Why not read these next?

Artificial Intelligence A.K.A the Robotic Takeover: A #TechTalk About Artificial Intelligence

First Case of Corona Virus in Nigeria! What can we learn from China’s Response