Python Programming for Biology: Bioinformatics and Beyond

10

See

Chapter 10

.

11

Fisher, R.A. (1936). The use of multiple measurements in taxonomic problems. Annals

of Eugenics 7(2): 179–188.

---

24

Machine learning

Contents

A guide to machine learning

Supervised machine learning

Unsupervised machine learning

Machine learning algorithms

Feature space

k-nearest neighbours

Distance between feature vectors

k-nearest neighbours in Python

Self-organising maps

A Kohonen map in Python

Feed-forward artificial neural networks

Training a neural network by back propagation

A Python neural network

A neural network for biological sequences

Support vector machines

A Python support vector machine

Support vector machine predictions

A guide to machine learning

When using computers to solve scientific problems there can be situations where you have

some  measured  data  and  a  related  property  of  the  data,  but  there  is  no  known  or  fixed

formula to link the two. Sometimes the link between the two sets of data may be easy for a

human to see, but otherwise difficult to encode in a computer algorithm. A simple example

of this would be in the reading of handwriting; humans do not write in a fixed typeface,

every letter of a given kind will be written slightly differently, and yet we can read most

other  people’s  handwriting  without  much  effort.  When  we  look  at  writing  we  attempt  to

recognise the letters and words, and where there is ambiguity we can use our intelligence

to infer what was intended by using the context of what the writing means, or any other

clues that we can glean. Writing a computer program to read handwriting is difficult, and

not nearly as reliable as a person would be. Nevertheless it can be done, and is put to good

---

use in the mechanised sorting of mail by postal (zip) code. The common trick to getting a

computer to perform tasks like this is not to program it with a designed and elaborate rule,

but rather to bestow a computer program with a degree of artificial intelligence so that it

can come up with its own rules and learn. The exercise whereby a program comes up with

its  own  rules  to  solve  a  problem  is  often  referred  to  as  machine  learning.  It  should  be

noted,  however,  that  we  usually  don’t  expect  a  computer  to  learn  a  task  perfectly;  if

perfection were possible we generally wouldn’t have to resort to such means. Instead it is

best  to  think  of  machine  learning  algorithms  as  making  predictions,  and  as  such  the

predictive power should be tested before we make reliance upon it. There are two kinds of

machine  learning  which  are  commonly  discussed,  supervised  learning  and  unsupervised

learning, and we will give examples of both in this chapter.