Machine Learning and Drug Discovery

At the moment I am working on a project to show some of what selenium can do. To be honest not something I have worked with before and my knowledge of HTML etc. is a good 10 years old so taking a while to get into it. I can open a webpage and I have managed to copy a table, just working out how to format that table! 

Drug Discovery:

Given my experience working in cancer research and related areas I was looking for some interesting articles on machine learning and found one entitled: 

Machine learning leads to breakthrough in drug discovery

Of course this immediately piqued my interested and I had a quick look. 

And this led me to: 

University of Warwick

It is very clever work and I cannot see a published article so there is not really much to go on at this point, but if they are really predicting interactions with a 99% success rate then that is an amazing achievement and is just another example that show the potential of machine learning. 

Machine learning to improve database performance?

I was doing some research the other day and came across an interesting article entitled.

A Machine Learning Approach to Databases Indexes - This is published by Google. 

As a database developer who has worked on massive databases where the biggest concern with going live with anything was to do with performance and now as a keen data scientist this sounds interesting. 

It will be interesting to see whether the neural networks that they have applied can be used in the general real world to improve performance to the extent that the have seen on their systems. Given the improvements that indexing can make to queries I would not be surprised to know that with modern systems there is a way to improve these systems further. 

As the paper is freely available I will just let anyone interested read it. 

Neural Networks

So I am not doing a proper post today because I am busy brushing up on my Neural Networks so that I can do some more details posts on this and hopefully how they can be implemented for many different uses. I have a couple of use cases at work that I will be able to use this for. 

The first is something which I logistic regression model could be used for but the new course that I am taking on Deep Learning has a model for detection of fraud using neural networks. The other is for the prediction of customer churn rate. 

The course can be found on Udemy, I am using this in conjunction with the Machine learning hands on book to increase my knowledge on neural networks and machine learning. 

Linear Regression

Linear regression is supposed to be the starting point of machine learning, it is the algorithm that most people start with and it is really easy to understand and super easy to implement in python. I won't go into details on the theory but in the most basic sense you plot points on a graph using known data and then extrapolate a prediction from it. 

So in my example I have some running data, I have a group of results for runners over a flat 5k route and a set of results over a hilly 3k route and I want to predict how fast any person is going to run the 3k route based on their 5k data. Very simple as I have one input variable against which I am going to make the prediction. The code below has comments showing how this is done with python.... 

Some helpful links here and here

The code: