In this episode Andrew and Danny run a simple demo that builds a Machine Learning Model using Azure Machine Learning and Python.
One of the most common questions we are asked about Machine Learning and AI is this: “Just how hard is it to use AI in my business?”
Amazingly Simple Code
We decided that we’d fire up one of the standard Machine Learning demos – anomaly detection in sensor data – and use this to demonstrate how we go about building, training and testing a simple machine learning model.
The code to build the model is only 3 lines!
To be fair, this is because we’re using pre-built libraries that do all the hard work for us. That’s the power of software – do that hard stuff once and its reuse is unlimited.
Where the Hard Parts Are
OK, so this demo is simple – deliberately so – because we wanted to highlight how easy the software engineering side of model building is.
There are two hard parts you need to look at:
- Choosing the right model
A big part of a Data Scientist’s role will be to curate data sets that contain the right mix of scenarios (input data + expected output) in order to create an accurate and unbiased mode.
Secondly, the Data Scientist will need to specify the correct type of model for the job in hand. Some scenarios will need regression models, some will need neural networks. In our case we chose a Decision Tree Classifier. Getting the model right has a huge impact on your results.
Watch the Video
What a year it’s been! Thinking back to January, we were dealing with the impact of the impending IR35 rule change. As with all software
Before we jump into creating an AI proof of concept for our clients, we always ask them a series of questions so that they understand
How do you cut the most pairs of shoes out of a hide of leather? And make sure there are no blemishes? All this and
How will AI ensure you don’t get a dented can of baked beans? Or that you don’t get green crisps in your packet of cheese