Created
October 5, 2017 07:11
-
-
Save selva86/133f36b7f82a3cdcc9f6a96dc555d1d3 to your computer and use it in GitHub Desktop.
Reproducible example for ks_plot
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
library(InformationValue) | |
library(ggplot2) | |
# 1. Import dataset | |
trainData <- read.csv('https://raw.githubusercontent.com/selva86/datasets/master/breastcancer_training.csv') | |
testData <- read.csv('https://raw.githubusercontent.com/selva86/datasets/master/breastcancer_test.csv') | |
# 2. Build Logistic Model | |
logitmod <- glm(Class ~ Cl.thickness + Cell.size + Cell.shape, family = "binomial", data=trainData) | |
# 3. Predict on testData | |
pred <- predict(logitmod, newdata = testData, type = "response") | |
# 4. If p > .5, then Class is 1 else 0 | |
y_pred <- ifelse(pred > 0.5, 1, 0) | |
y_act <- testData$Class | |
# 5. Accuracy | |
mean(y_pred == y_act) # 94% | |
ks_table <- InformationValue:::ks_table | |
ks_plot <- function (actuals, predictedScores) { | |
rank <- 0:10 | |
ks_table_out <- ks_table(actuals = actuals, predictedScores = predictedScores) | |
perc_positive <- c(0, ks_table_out$cum_perc_responders) * 100 | |
perc_negative <- c(0, ks_table_out$cum_perc_non_responders) * 100 | |
random_prediction <- seq(0, 100, 10) | |
df <- data.frame(rank, random_prediction, perc_positive, perc_negative) | |
df_stack <- stack(df, c(random_prediction, perc_positive, perc_negative)) | |
df_stack$rank <- rep(rank, 3) | |
df_stack$delta <- df_stack$values[12:22] - df_stack$values[1:11] | |
values <- df_stack$values | |
ind <- df_stack$ind | |
rowmax <- which.max(ks_table_out$difference) | |
l_start <- ks_table_out[rowmax, "cum_perc_non_responders"] | |
l_end <- ks_table_out[rowmax, "cum_perc_responders"] | |
print(ggplot2::ggplot(df_stack, aes(x = rank, y = values, | |
colour = ind, label = paste0(round(values, 2), "%"))) + | |
geom_line(size = 1.25) + | |
labs(x = "rank", y = "Percentage +Ve & -Ve Captured", | |
title = "KS Chart", subtitle=paste("KS Statistic: ", ks_stat(actuals, predictedScores))) + | |
theme(plot.title = element_text(size = 20, | |
face = "bold")) + | |
geom_text(aes(y = values + 4)) + | |
scale_x_continuous(breaks=0:10, labels=0:10) + | |
geom_segment(x = rowmax, y = l_start*100, xend = rowmax, yend = l_end*100, col="red", arrow = arrow(length = unit(0.05, "npc"), ends="both"), linetype = "dashed", lwd=1)) | |
} | |
ks_plot(y_act, y_pred) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment