Back in April I started the Machine learning MOOC of Coursera with the idea of getting to know the basics. I wasn’t certain about whether I could complete the course or not, without even having started my computer science university degree yet.
The main topics covered by this MOOC are:
Preventing overfitting, easing treatment of new examples
A layered structure for learning, inspired by the biological neural networks
Clustering, grouping unlabeled data
It also teaches about practical uses of machine learning:
Support Vector Machines
Algorithm for classification
Detecting anomalies in a large data set
Recommending to users things they may like
Photo OCR (Optical Character Recognition)
Identifying and recognizing objects, words, and digits in an image
It has been achievable. Nearly everything has been math; mainly linear algebra. Concerning the programming assignments, they were challenging: they consisted of applying math to programs so that they performed the appropriate machine learning processes. They had to be done in MATLAB or Octave.
More importantly, I feel that I’ve learnt many useful concepts. I’ve also found this field of study very interesting. I’d be so happy to be engaged in more learning and experience on this field. Thus I’m already messing with data science projects.
This MOOC was created by Stanford University and taught by Andrew Ng, co-founder of Coursera.