This book covers common aspects in predictive modeling:

• A. Data Preparation / Data Profiling
• B. Selecting best variables (dataviz)
• C. Assessing model performance
• D. Miscellaneous
  

And it is heavly based on the funModeling package from the R language . Please install before starting :)

install.packages("funModeling")

• Model creation consumes around 10% of almost any predictive modeling project; funModeling will try to cover remaining 90%.
• It's not only the function itself, but the explanation of how to interpret results. This brings a deeper understanding of what is being done, boosting the freedom to use that knowledge in other situations regardless of the language.
  

Hopefully this book barerly has an end, it will be updated periodically. Next planned chapter is a case study in predictive modeling. And you can contribute! below the github link.

This chapter will cover three types of plots which aim to understand what are the most correlated numeric variables against a target variable.

• Overview:
  • Analysis purpose: To identify if the input variable is a good/bad predictor through visual analysis.
  • General purpose: To explain the decision of including -or not- a variable to a model to a non-analyst person.
    
  
  

Constraint: Target variable must contain only 2 values. If it has NAvalues, they will be removed.

Last two plots have the same data source, showing the distribution ofhas_heart_disease in terms of gender. The one on the left shows in percentage value, while the one on the right shows in absolute value.

Gender variable seems to be a good predictor, since the likelihood of having heart disease is different given the female/male groups. it gives an order to the data.

• The likelihood of having heart disease for males is 55.3%, while for females is: 25.8%.
• The heart disease rate for males doubles the rate for females (55.3 vs 25.8, respectively).