AI and Machine Learning for Qlik Developers - Part 2

This is the second article on my project to create an AI only using Qlik Script.

First article can be found here. Part 3 can be found here.

Starting easy

The first prototype will be completely pointless in terms of what it can do for us, but my purpose is to slowly introduce Machine Learning terminology and methods using only Qlik Script.

Required Qlik Script skills

For the first example, I will use the following very basic Qlik Script functions and operators, that most Qlik Delvelopers probably anyway are using on a daily basis. If you are not familiar with some of them, now is the time to go to the Qlik community.

Set, Let, Load, Inline, RowNo(), NoOfRows(), Peek(), For...to..next, {concept of dollar expansion}, +-/*, and...or...

For the little bit more complicated prototypes later on in this project, you also need to know the following functions:

Round(), Rand(), AutoGenerate(), Sqr(), Pow(), Resident, Sub, Call, Concatenate, NoOfRows(), if..then..else, exit, switch...case

If you know them already, then you can build an AI!

Concept of the first prototype

We will create a Qlik script that will teach the computer to predict the output of a very simple mathematical expression without knowing the original formula.

Take a closer look at this data:

The formula that created this data looks like this: 2x+y=Output

This formula is what we will train our AI to figure out. In this very simple first prototype we will tell our AI that it should look for a formula of this type: Ax+By=Output, and the task is to find out the correct values of A and B. This is of course helping our AI a lot, and arguably we could say that this is therefore not a "real" AI.... but my point in this first prototype is just to explain the basic machine learning terminology.

The process of helping the AI find a formula based on given input and output is usually referred to as "Supervised Learning". "Unsupervised Learning" is when output data is not existing at all and the AI need to find a pattern of the input data completely by itself.

Later in this series we will let our AI figure out more advanced things... but we will start simple, and in the next article we will look at actual script examples of this first prototype.

Neural Networks and Neurons

Even if we start with this simple prototype you will get a little bit familiar with the concept of Supervised and Nonsupervised training, Classification and Regression, Neurons, Neural Networks, Different kinds of Layers.

?

Thank you for reading. Please share if you like it, and comment on the post if you have questions or feedback about the content.

Connect with me here on LinkedIn if you want to be sure to see the end result, and eventually get a copy of the complete Qlik Script.