Data 605 Discussion 11

I’m using ChemicalManufacturingProcess from the package AppliedPredictiveModeling.

Chemical Manufacturing Process Data

Description

This data set contains information about a chemical manufacturing process, in which the goal is to understand the relationship between the process and the resulting final product yield. Raw material in this process is put through a sequence of 27 steps to generate the final pharmaceutical product. The starting material is generated from a biological unit and has a range of quality and characteristics. The objective in this project was to develop a model to predict percent yield of the manufacturing process. The data set consisted of 177 samples of biological material for which 57 characteristics were measured. Of the 57 characteristics, there were 12 measurements of the biological starting material, and 45 measurements of the manufacturing process. The process variables included measurements such as temperature, drying time, washing time, and concentrations of by–products at various steps. Some of the process measurements can be controlled, while others are observed. Predictors are continuous, count, categorical; some are correlated, and some contain missing values. Samples are not independent because sets of samples come from the same batch of biological starting material.

Loading the data…

data(ChemicalManufacturingProcess)

Used bagImpute method to impute the missing values.

bagImpute <- preProcess(ChemicalManufacturingProcess[,-c(1)], method=c('bagImpute'))

CMP <- predict(bagImpute, ChemicalManufacturingProcess)
visdat::vis_miss(CMP) +
  theme(axis.text.x = element_text(color = "grey20", size = 4, angle = 45, hjust = 0, vjust = .8, face = "plain"))

Here are my steps to create a linear model using method = ‘lm’

Step 0: Set a seed

# Make this reproducible
set.seed(45)

Step 1: Removing the zero variance variables from the dataset CMP

CMP_2 <- CMP %>% select(-one_of(nearZeroVar(., names = TRUE)))

Step 2: Create train and test split

train <- createDataPartition(CMP_2$Yield, times = 1, p = 0.8, list = FALSE)
CMP2_train_df <- CMP_2[train, ]
CMP2_test_df <- CMP_2[-train, ]

Step 3: Train the model

lm_model <- train(
                  Yield ~ ., data = CMP2_train_df,
                  method = "lm",
                  tuneLength = 24,
                  trControl = trainControl(method = "repeatedcv", repeats = 8),
                  preProcess=c('center', 'scale')
)
## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading

Step 4: Examining model results

lm_model$results %>%
  select(Rsquared)
##    Rsquared
## 1 0.4834498

The R2 of this linear model on the train set is 51.92%

Prediction Results on the test set

# Make predictions
pred <- predict(lm_model, CMP2_test_df)
## Warning in predict.lm(modelFit, newdata): prediction from a rank-deficient fit
## may be misleading
# Model performance metrics
postResample(pred=pred, obs=CMP2_test_df$Yield)
##      RMSE  Rsquared       MAE 
## 1.6214377 0.3982772 1.2849060

The R2 of prediction results on the test set dropped to 35.25%.

Checking on assumptions

checking assumptions automatically…

fit <- lm(Yield ~ ., CMP2_train_df)

try(gvlma.lm(fit))
## Error in solve.default(sigwhat) : 
##   system is computationally singular: reciprocal condition number = 1.70266e-22

The assumptions failed automatically right there. The error message just means my design matrix is not invertible and therefore can’t be used to develop a regression model. This results from linearly dependent columns, i.e. strongly correlated variables. Examine the pairwise covariance (or correlation) of your variables to investigate if there are any variables that can potentially be removed. You’re looking for covariances (or correlations) >> 0. Alternatively, you can probably automate this variable selection by using a forward stepwise regression.

Further Research

Using grepl I was able to find the element position within the list results of names(getModelInfo()). Then I found the method lmStepAIC might be the way to do a variable selection to eliminate variables that are deemed not useful in passing the assumption check.

lst <- names(getModelInfo())
lst_r <- grepl("^lm",lst)
lst[lst_r]
## [1] "lm"        "lmStepAIC"