Machine Learning

According to Tom Mitchell, what is the requirement for the computer system to learn

• Data set

• Task

• Performance measure

• Knowledge

Machine learning is mostly used when

• Human expertise doesn't exist

• Model must be customised to personal need

• Model use huge amount of data

• Interpreting the insight from descriptive data such as mean, median etc

Select a possible application of the machine learning

• Fraud detection

• Recognising a happy face

• Differentiate X-ray image

• Recognise your lecturer's hand writing

To recognise the face as an object, the image of face is breakdown into combination of edges, edges, and pixel.

• True

• False

What is the implication of "false positive" increase to the precision measure

• increase precision

• decrease precision

• doesn't has any effect to precision

What is the implication of decreasing "false negative" to the recall measure

• decreasing recall

• increasing recall

• doesn't have any effect to recall

In the case of pregnancy prediction, if the predictor output is always "pregnant" what is the recall value?

• 1

• 0

• 0.1

• Undefined

In the case of pregnancy prediction, if the predictor output is always "not pregnant" what is the precision value?

• 1

• 0

• 0.5

• Undefined

The unsupervised learning is better used for prediction

• True

• False

Artificial neural network consists of following fundamental components

• Neurons as nodes

• Synapses as weight

• Neurons as weight

• Synapses as nodes