- Extreme Optimization
- Documentation
- Statistics Library User's Guide
- Statistical Variables
- Continuous Variables
- Categorical Variables
- Variable Collections
- General Linear Models
- Regression Analysis
- Analysis of Variance
- Time Series Analysis
- Multivariate Analysis
- Continuous Distributions
- Discrete Distributions
- Multivariate Distributions
- Hypothesis Tests
- Histograms
- Random Numbers
- Appendices

- Continuous Variables
- Numerical Variables

Numerical Variables | Extreme Optimization Numerical Libraries for .NET Professional |

Variables whose observations are numeric in nature are called numerical variables. In the **Extreme Optimization
Numerical Libraries for .NET**, numerical variables are implemented by the NumericalVariable class.

Numerical variables can be constructed in a variety of ways. The NumericalVariable class has six constructors that come in three groups.

The first group uses a Double array as the source of the data. The first variant has two parameters. The first is a string that specifies the name of the variable. The second parameter is a Double array. The second variant only takes one parameter: a Double array containing the data values.

double[] dataArray = new double[] {62, 77, 61, 94, 75, 82, 86, 83, 64, 84, 68, 82, 72, 71, 85, 66, 61, 79, 81, 73}; NumericalVariable variable1 = new NumericalVariable(dataArray); NumericalVariable variable2 = new NumericalVariable("Data", dataArray);

The second group uses a Vector

Vector dataVector = Vector.Create(dataArray); NumericalVariable variable3 = new NumericalVariable(dataVector); NumericalVariable variable4 = new NumericalVariable("Data", dataVector);

The third group of constructors uses a System.Data

DataColumn column; // Connect to a data source and retrieve the column from a DataTable NumericalVariable variable5 = new NumericalVariable(column); NumericalVariable variable6 = new NumericalVariable("Data", column);

A number of static (Shared in Visual Basic) methods create numerical variables that span a range of numbers. The CreateRange method is overloaded. The first overload takes one integer argument that specifies a maximum value. It returns a NumericalVariable whose observations are the integers from 0 up to but not including the maximum value. The second overload takes two integer arguments that specify the minimum and maximum values. Once again, the maximum value is not included.

The third overload takes three arguments. The first is the number of observations. The second and third arguments are real numbers that specify the lowest and highest value. This method returns a NumericalVariable containing the specified number of observations that are equally spaced between the lowest and highest value. In this case, the highest value is included.

The CreateLogarithmicRange method also takes three arguments. The first is the number of observations. The second and third arguments are real numbers that specify the lowest and highest value. This method returns a NumericalVariable containing the specified number of observations whose logarithms are equally spaced between the lowest and highest value. This means that the ratio between two successive observations is a constant. The highest value is included.

The following example creates two variables with 5 observations. The first contains the first five multiples of 1000. The second has a logarithmic scale and contains the first 5 powers of 10, starting with 0:

NumericalVariable variable7 = NumericalVariable.CreateRange(5, 1000, 5000); NumericalVariable variable8 = NumericalVariable.CreateLogarithmicRange(5, 1, 10000);

In addition, variables can be created by VariableCollection objects, by performing arithmetic operations on them (see below), and several other means.

Numerical variables have the widest range of descriptive statistics available. The values are calculated as needed, and cached. For some values, the calculation may be lengthy. In this case, the value is returned by a method instead of a property. The following tables list the descriptive statistics that are available for numerical variables:

The purpose of a measure of location is to provide a typical or central value to describe the data.

The median is the middle value of a sorted list of observations. If a variable has an even number of observations, then the median is the average of the two middle values.

Method | Description |
---|---|

Returns the geometric mean of all observations. | |

Returns the harmonic mean of all observations. | |

Returns the mean of the middle 50% of observations. | |

Returns the mean of the observations after eliminating the specified percentage of extreme values. | |

Returns the mean of the observations after setting the specified percentage of extreme values to the lowest or highest value. |

The mid-mean is a special case of the trimmed mean. Providing a value of 100 for the percentage to the GetTrimmedMean method returns the median. The Winsorized mean is similar to the trimmed mean, but instead of eliminating the extreme values, they are set to the lowest or highest value. For example, for the 10% Winsorized mean, the 5% smallest values are set to equal the value at the 5% percentile, while the 5% largest values are set to equal the value at the 95% percentile.

The examle below shows how to use some of the most common measures of location:

Console.WriteLine("Mean: {0:F1}", variable1.Mean); Console.WriteLine("Median: {0:F1}", variable1.Median); Console.WriteLine("Trimmed Mean: {0:F1}", variable1.GetTrimmedMean(10)); Console.WriteLine("Harmonic Mean: {0:F1}", variable1.GetHarmonicMean()); Console.WriteLine("Geometric Mean: {0:F1}", variable1.GetGeometricMean());

Measures of scale are used to characterize the spread or variability of a data set.

Property | Description |
---|---|

Returns the unbiased variance of the data. | |

Returns the variance of the data. | |

Returns the unbiased standard deviation of the data. | |

Returns the standard deviation of the data. | |

Returns the root-mean-square. | |

Returns the difference between the largest and the smallest value. | |

Returns the smallest value. | |

Returns the largest value. |

The variance and the standard deviation are always the unbiased versions. To get the biased (population) standard deviation, use the RootMeanSquare property.

Method | Description |
---|---|

Returns the average absolute deviation from the mean. | |

Returns the median of the absolute deviation from the mean. | |

Returns the difference between the first and the third quartile. |

The average absolute deviation is the mean of the absolute difference between each value and the mean. Because it does not square the distance from the mean, it is less affected by extreme values. The median absolute deviation is the median of the absolute difference between each value and the mean. It is even less influenced by extreme values because the median is less affected by extreme values than the mean.

The inter-quartile range is the difference between the 75% and the 25% percentile values. It is a measure of the variability of values close to the mean.

The examle below shows some of these properties and methods:

Console.WriteLine("Standard deviation: {0:F1}", variable1.StandardDeviation); Console.WriteLine("Variance: {0:F1}", variable1.Variance); Console.WriteLine("Range: {0:F1}", variable1.Range); Console.WriteLine("Inter-quartile range:{0:F1}", variable1.GetInterQuartileRange());

The remaining properties cover the higher moments (skewness and curtosis) and the raw sums:

Property | Description |
---|---|

Returns the unbiased skewness of the data. | |

Returns the skewness of the data. | |

Returns the unbiased kurtosis supplement of the data. | |

Returns the kurtosis supplement of the data. | |

Returns the sum of all the elements. | |

Returns the sum of the squares. |

The skewness is a measure for the lack of symmetry of the distribution of a variable. The kurtosis is a measure of
the peakedness compared to the normal distribution. The Kurtosis property returns the kurtosis
*supplement*, which is the difference between the 'real' kurtosis and the kurtosis of the normal distribution,
which equals 3.

Several measures of correlation are available. The GetCovariance method returns the covariance between two variables. The GetCorrelation method returns the Pearson correlation between two variables. The GetRankCorrelation method returns the Spearman rank correlation between two variables.

Each of these methods has two variants. The first is an instance method that takes the second variable as its only argument. It compares the current instance with its argument. The second variant is a static (Shared in Visual Basic) method that takes two variables as arguments and computes the value for its two arguments.

The GetAutocorrelation method returns the Pearson correlation of a variable with itself. An optional integer argument specifies the lag.

By default, missing values have the value Double.NaN. You can change this by setting the variable's MissingValue property.

Missing values are ignored during the calculation of descriptive statistics. To force a different behavior, you must call the ReplaceMissingValues method. This method takes one or two parameters. The first is a MissingValueAction value that determines the action that is to be taken when a missing value is encountered. The options are summarized in the table below. The second, optional parameter is a replacement value, if one is required. It defaults to zero.

Member Name | Description |
---|---|

Missing values are ignored. | |

All missing observations are discarded. | |

Missing values are ignored. | |

Missing values are replaced with the value of the previous observation. If the first observation is missing, it is replaced with a user-specified value, or 0. | |

Missing values are replaced with the value of the next observation. If the last observation is missing, it is replaced with a user-specified value, or 0. | |

Missing values are replaced with a user-specified value, or 0. | |

A MissingValueException is thrown. |

The IsMissing method indicates whether an observation is missing. Its only parameter is the index of the observation.

Numerical variables can be combined using arithmetic operations to form new variables. The result of such an operation is a variable whose components equal the operation applied to the corresponding components of the operand(s).

For languages that support operator overloading, the arithmetic operators, +, -, *, / have been overloaded. For languages that don't support operator overloading, static (Shared in Visual Basic) methods are provided.

The operations are summarized in the following table:

Operator | Static method | Description |
---|---|---|

+x | (no equivalent) | Returns the variable x. |

-x | Negate(x) | Returns the negation of the variable x. |

x + y | Add(x, y) | Adds the variables x and y. |

x + a | Add(x, a) | Adds the variable x and the real number a. |

a + x | Add(a, x) | Adds the real number a to the variable x. |

x - y | Subtract(x, y) | Subtracts the variables x and y. |

x - a | Subtract(x, a) | Subtracts the real number a from the variable x. |

a - x | Subtract(a, x) | Subtracts the variable x from the real number a. |

x * y | Multiply(x, y) | Multiplies the variables x and y. |

x * a | Multiply(x, a) | Multiplies the variable x and the real number a. |

a * x | Multiply(a, x) | Multiplies the real number a and the variable x. |

x / y | Divide(x, y) | Divides the variable x by y. |

x / a | Divide(x, a) | Divides the variable x by the real number a. |

a / x | Divide(a, x) | Divides the real number a by the variable x. |

- | Power(x, a) | Raises the variable's observations to the power a. |

- | Max(x, y) | Selects the largest observation from the variables x and y. |

- | Min(x, y) | Selects the smallest observation from the variables x and y. |

The NumericalVariable class has a number of additional methods that may be useful. The Normalize method returns the variable rescaled to have a mean of zero and a standard deviation equal to one. The Apply method lets you apply any real function with one argument to each value of the variable. The result is a new variable. You can use this method to transform variables in preparation for a regression analysis.

The following example prepares data for a fit of the function

*y* = *ae ^{bx1+cx2x3}*.

The dependent variable is transformed using the Math.Log method. A new variable, *z*, is
created to hold the product of *x _{2}* and

*x*. This transforms the above function into a form suitable for multiple linear regression:

_{3}log *y* = log *a + bx _{1} + cz.*

NumericalVariable Y, X1, X2, X3; // ... NumericalVariable logY = Y.Apply(new RealFunction(Math.Log)); NumericalVariable Z = X2 * X3;

Note that, because Visual Basic .NET (2003) does not support operator overloading, the shared Multiply method must be used.

The Sort method sorts the observations. The order is ascending by default, but can be specified by passing a parameter of type SortOrder.

Copyright Â© 2004-2015,
Extreme Optimization. All rights reserved.

*Extreme Optimization,* *Complexity made simple*, *M#*, and *M
Sharp* are trademarks of ExoAnalytics Inc.

*Microsoft*, *Visual C#, Visual Basic, Visual Studio*, *Visual
Studio.NET*, and the *Optimized for Visual Studio* logo

are
registered trademarks of Microsoft Corporation.