woeBal: Weight of Evidence Balance Sheet

In ralmond/CPTtools: Tools for Creating Conditional Probability Tables

Weight of Evidence Balance Sheet

Description

Creates a weight of evidence balance sheet from a history of marginal distributions.

Usage

woeBal(hist, pos, neg, obs=NULL, title = "Evidence Balance Sheet",
       col = rev(colorspread("slategray",ncol(hist),maxsat=TRUE)),
       posCol="cyan", negCol="red", stripCol=c("white","lightgray"),
       lcex = 0.65)

Arguments

hist	A matrix whose rows represent time points (after tests) and columns represent probabilities.
pos	An expression for selecting rows of the cpf which corresponds to the hypothesis.
neg	An expression for selecting the rows corresponding to the complement of the hypothesis. (The default value is -pos if pos is numeric; !pos if pos is logical, and setdiff(rownames(cpf),pos if pos is a character vector.
obs	An optional character vector of the same length as the number of rows of hist giving the value observed at each step
title	Title for plot
col	A list of color values for probability bars.
posCol	The color to be used for bars showing positive weights of evidence.
negCol	The color to be used for bars showing negative weights of evidence.
stripCol	The colors to be used for the time step labels. Setting this to a vector of two colors creates alternate color stripes. Set this to "white" to disable that effect.
lcex	Character expansion size for labels.

Details

This constructs a weight of evidence balance sheet (Madigan, Mosurski, and Almond, 1997) showing the changes to the probability distribution and weight of evidence for each change in the probability. The probabilities are given in the hist argument in which each row should be a probability distribution for the target variable. The labels for the plot are taken from the row labels of the hist argument.

Madigan, Mosurski and Almond (1997) note that the definition of weight of evidence is somewhat problematic if the hypothesis variable is not binary. In that case, they recommend partitioning the states into a positive and negative set. The pos and neg are meant to describe that partition. They can be any expression suitable for selecting columns from the hist matrix. This function calls woeHist() to calculate weights of evidence.

The row names of hist are printed left-justified in the leftmost column. If observed values (obs) are supplied, they are printed right justified in the same column.

Value

The midpoints of the bars (see barplot) are returned invisibly.

Side Effects

Starts a new plotting page and creates three side-by-side plots, one for the labels, one for the probability bars and one for the weight of evidence bars.

Author(s)

Russell Almond

References

Good, I. (1971) The probabilistic explication of information, evidence, surprise, causality, explanation and utility. In Proceedings of a Symposium on the Foundations of Statistical Inference. Holt, Rinehart and Winston, 108-141.

Madigan, D., Mosurski, K. and Almond, R. (1997) Graphical explanation in belief networks. Journal of Computational Graphics and Statistics, 6, 160-181.

Almond, R. G., Kim, Y. J., Shute, V. J. and Ventura, M. (2013). Debugging the Evidence Chain. In Almond, R. G. and Mengshoel, O. (Eds.) Proceedings of the 2013 UAI Application Workshops: Big Data meet Complex Models and Models for Spatial, Temporal and Network Data (UAI2013AW), 1-10. http://ceur-ws.org/Vol-1024/paper-01.pdf

Almond, R.G., Mislevy, R.J., Steinberg, L.S., Williamson, D.M. and Yan, D. (2015) Bayesian Networks in Educational Assessment. Springer. Chapter 7.

See Also

readHistory, woeHist, barplot, Colors

Examples

sampleSequence <- read.csv(system.file("testFiles","SampleStudent.csv",
                                       package="CPTtools"),
                           header=TRUE,row.names=1)

woeBal(sampleSequence[,c("H","M","L")],c("H"),c("M","L"),lcex=1.25)
woeBal(sampleSequence[,c("H","M","L")],c("H"),c("M","L"),
      obs=sampleSequence[,"Acc"],lcex=1.25)

ralmond/CPTtools documentation built on Dec. 27, 2024, 7:15 a.m.