Publications by Gary Hutson
RODBC helper function
The number of times I have to connect to SQL and I forget part of the RODBC command to connect to an internal data table. As part of a project I am working on I have been connecting to lots of different sources and became tired of typing lots of lines and repeating the same command. Therefore, I have been rescued by R’s lovely functions. The fu...
2535 sym R (474 sym/2 pcs)
Machine Learning Training – Draper and Dash Healthcare Predictive Analytics – Summer Discount
There is an offer on with my company Draper & Dash to get a discount on ML training for your organisation. Contact the sales team to find out more about this training opportunity. Plus you get to meet our great data science team at Draper and Dash. The below is me and colleagues in action at the event we held at St Pancreas Hotel and Conference...
771 sym
Confusion matrices – evaluating your classification models
As part of my companies commitment to allow our users access to some of our code – we have created a visual way you can assess your accuracy in a confusion matrix. The below was posted on our blog site and shows how to interpret a confusion matrix. Confusion Matrices – Evaluating your classification models I hope the attached article is us...
938 sym 2 img
Deploying a ML model in R
As a product of a series of training sessions Draper & Dash have been undertaking on Machine Learning using R – a candidate from the course asked “how do you split a model to later be used in a live / production environment?”. This post aims to answer that question: Deploying a Trained Supervised ML Model I hope you find this useful and ...
795 sym
Bucketing and highlighting dominant predictors in your ML models
Our company have been at it again. This time from my colleagues and fellow Data Scientist Alfonso Portabales. For this post we look at Draper and Dash’s custom method of highlighting dominant predictors in your machine learning models: Bucketing and highlighting dominant predictors in your ML models If you are interested in our solutions in ...
1288 sym
Predictive Solutions Series – Draper & Dash’s Stranded and Super Stranded Patient Module
Lots of exciting things are happening with Draper and Dash at the moment. We have been working with key healthcare partners to design some core predictive machine learning algorithms to enable trusts to more effectively manage their performance, demand and capacity pressures. This series focuses on Stranded Patients and Super Stranded patients �...
2403 sym 2 img
Getting on the meet-up bandwagon – our first meet up event
My company Draper and Dash have tasked me with organising a wider meet-up event for anyone who is interested in AI / ML in healthcare. This wider working group consists of people from different sectors, however they are interested in how we can apply AI / ML methods in their organisations. Why did we choose meet-up Meet-up seems like a great way...
1527 sym 24 img
Command Centre amplification with predictive analytics and machine learning
Recently, our team at Draper and Dash have been busy creating an NHS operational command centre. This command centre is different, as it uses a collection and ensemble of cutting edge predictive and machine learning techniques. To read the blog you can access this below: Command Centre amplification with predictive analytics and machine learning...
1150 sym
OddsPlotty – the first official package I have ‘officially’ launched
Motivation for this The background to this package linked to a project I undertook about a year ago. The video relates to the project and the how R really sped up the process. The exam question was to use a regression model to predict admissions and we had to evaluate, as a consequence, 60 different variations. In Excel, this would have been a ...
3575 sym R (89 sym/1 pcs) 8 img
Threading and Caret: Burning your CPU to improve model training speed
When I am doing a Machine Learning project with R it is crucial to save those precious seconds, minutes, hours in model training. To make sure you optimise your CPU and crank up the performance you will need to load in the parallel and doParallel libraries, alongside library(caret). A wrapper function to help with this I have created a wrapper fu...
3766 sym R (550 sym/4 pcs) 2 img