Problem statement

The aim of this project is to process random series of data representing sensor data.

Outliers have to be detected. Times are expressed in hours with decimals. Series of

data are not regularly distributed but covers 24 hours. Here, we will used relative time: it is worth 0 at the beginning of the day.

Outliers

Contextual outliers are defined as data points, which, in the contact of previous and

future data points, seem highly improbable. These points have been detected as the

ones that follow the equations simultaneously:

All the data points that satisfy the above equation are outliers.

It is requested to:

1. create a class Serie that collects doubles representing either relative times

in hours or measurement data. Each serie should be named and must contain

methods to return its name, its size and its values as double[]

2. design a toString() method to display the series

3. with Serie, generate a random serie of 100 (start by 10 for debugging)

consecutive values of CO2 between 400 and 2500 using a method void !

,! populateWithValues(int numberOfValues, double min, double max) for populating the

serie.

4. with Serie, generate a non equaly distributed serie of 100 (start by 10 for

debugging) consecutive relative times in hours covering 24 hours using a

method void populateWithTimes(int numberOfValues, double max) for populating the

serie.

5. add methods to compute average and standard deviation.

optional detect the outliers from the random data. If no outliers is found, reduce the

value of for testing.

6.2 Clues

Math.random() generates a random number between 0 and 1.

Math.abs(doubleValue) computes the absolute value of doubleValue.

use the method set to replace a value in an ArrayList.

to sort an ArrayList, you can use the method sort bound to the class java. !

,! util.Collections.

to copy the elements of an ArrayList of type T, use the method toArray( !

,! T[] arrayWhereDataAreCopied)

pdiff. fdiffk