LF.data is simulated case control data where the ratio of cases:controls is 1:1. The dataset contains 200 observations and 50 binary predictors. Columns 1,2,....,50 are the binary predictors. Column 51 is the true binary response and column 52 is a noisy binary response. Each predictor is simulated as an independent Bernoulli random variable with success probability 0.5. The response variable is simulated from the model Y=(X4 and X5) or (X9 and X10)
The true response in column 51 takes value 1 if either of these conditions occurs and takes value 0 otherwise. Column 52 represents the response with some error. The probability that the noisy response (column 52) has the same value as the true response is 90