Updated on October 31, 2015

## Gradient Descent Optimization

In many machine learning algorithm, the goal is to find a function or parameters that allows us to approximate or modelize unknown observable data. Those data could come from device measurement, web crawling, empirical observations etc. Generally speaking we have …

Updated on October 18, 2015

## Hosting a Node.js application with Apache

Several months ago, I was very excited by the annouce of the establishment of one of the biggest if not the biggest data center in Canada near Montreal. In its offer, OVH proposes VPS at a very low entry price. …

Updated on October 18, 2015

## k-means

For a first article, we'll see an implementation in Matlab of the so-called k-means clustering algorithm. K-means algorithm is a very simple and intuitive unsupervised learning algorithm. Indeed, with supervised algorithms, the input samples under which the training is performed …

Updated on October 18, 2015

## Use MongoDB .NET driver with Matlab - Part.2

I've presented the basics to connect with MongoDB database server and populate a collection with simple documents in a previous post. We're now going a step further by showing how to retrieve data and play with them. Let's say …

Updated on October 18, 2015

## Use MongoDB .NET driver with Matlab

After an article on the use of SqlLite with Matlab, here is another way to interact with database, a NoSQL database powered by MongoDB.

First of all, you have to install MongoDB on your computer. For my part, I …

Updated on October 18, 2015

## Working with SQLite database in Matlab

You might have already wonder how you could deal with large amount of data taken in lab ? For example, you might have led some experiences with different conditions / hypothesis, performed measurements and store the resulting raw samples into …

Updated on October 18, 2015

## Combinatorial analysis - Enumerator

In mathematics, the branch that studies number arrangement is called combinatorial analysis. This a pretty tough exercice in general because one we might think of all possiblilities, all ways numbers can be arranged. For example, try to figure how to …

Updated on October 18, 2015

## 2-D logistic regression

We have previously presented a proposed implementation for the k-means algorithm on computed samples. We saw that the result performed quite well clustering our samples. The next step now is to evaluate the probablility that a new sample belongs to …