Statistics Registers

Implement a set of statistics registers in the style of a pocket calculator.

The available routines are:

def  Clear():                        clear the stats registers
def  Show():                         print the contents of the stats registers
def  Add(x, y):                      add an X,Y pair
def  Subtract(x, y):                 remove an X,Y pair
def  AddWeighted(x, y, z):           add an X,Y pair with weight Z
def  SubtractWeighted(x, y, z):      remove an X,Y pair with weight Z
def  Mean():                         arithmetic mean of X & Y
def  StdDev():                       standard deviation on X & Y
def  StdErr():                       standard error on X & Y
def  LinearRegression():             linear regression
def  LinearRegressionVariance():     est. errors of slope & intercept
def  LinearRegressionCorrelation():  the regression coefficient
def  CorrelationCoefficient():       relation of errors in slope & intercept

see:http://stattrek.com/AP-Statistics-1/Regression.aspx?Tutorial=Stat

pocket calculator Statistical Registers, Pete Jemian, 2003-Apr-18

Source code documentation

class jldesmear.jl_api.StatsReg.StatsRegClass

pocket calculator Statistical Registers class

Add(x, y)

add an X,Y pair to the statistics registers

Parameters:

• x (float) – value to accumulate
• y (float) – value to accumulate

AddWeighted(x, y, z)

add a weighted X,Y+/Z trio to the statistics registers

Parameters:

• x (float) – value to accumulate
• y (float) – value to accumulate
• z (float) – variance (weight = 1/z^2) of y

Clear()

clear the statistics registers: \(N=w=\sum{x}=\sum{x^2}=\sum{y}=\sum{y^2}=\sum{xy}=0\)

CorrelationCoefficient()

relation of errors in slope and intercept

Returns:correlation coefficient Return type:float

LinearEval(x)

Evaluate a linear fit at the given value: \(y = a + b x\)

Parameters:x (float) – independent value, x Returns:y Return type:float

LinearRegression()

For (x,y) data pairs added to the registers, fit and find (a,b) that satisfy:

\[y = a + b x\]

Returns:(a, b) for fit of y=a+b*x Return type:(float, float)

LinearRegressionCorrelation()

the regression coefficient

Returns:(corr_a, corr_b) – is this correct? Return type:(float, float) See:http://stattrek.com/AP-Statistics-1/Correlation.aspx?Tutorial=Stat Look at “Product-moment correlation coefficient”

LinearRegressionVariance()

est. errors of slope & intercept

Returns:(var_a, var_b) – is this correct? Return type:(float, float)

Mean()

arithmetic mean of X & Y

\[(1 / N) \sum_i^N x_i\]

Returns:mean X and Y values Return type:float

Show()

contents of the statistics registers

StdDev()

standard deviation on X & Y

Returns:standard deviation of mean X and Y values Return type:(float, float)

StdErr()

standard error on X & Y

Returns:standard error of mean X and Y values Return type:(float, float)

Subtract(x, y)

remove an X,Y pair from the statistics registers

Parameters:

• x (float) – value to remove
• y (float) – value to remove

SubtractWeighted(x, y, z)

remove a weighted X,Y+/-Z trio from the statistics registers

Parameters:

• x (float) – value to remove
• y (float) – value to remove
• z (float) – variance (weight = 1/z^2) of y

determinant()

compute and return the determinant of the square matrices:

|  sum_w   sum_x      |          |  sum_w   sum_y      |
|  sum_x   sum_(x^2)  |          |  sum_y   sum_(y^2)  |

Returns:determinants of x and y summation matrices Return type:(float, float)