Car Mileage Predictor

Socrates Krishnamurthy

Introduction

Calculation of 'Mileage' of various cars across the Globe is challenging and time consuming. This interactive application helps predict the mileage of a car when a set of parameters are provided.

What this application do -

Uses mtcars dataset provided by 'Motor Trends Magazine' as reference
Uses Regression Alogorithm (Linear Model)
Gets User input for an unknown car
Predicts Mileage of unknown car using above Linear Model

Dataset and Regression Model

library(datasets); data(mtcars)

linearModel <- lm(mpg ~  wt + cyl + disp + hp + am, data = mtcars)

Linear Model - Intercepts and Slope

linearModel[1]

$coefficients
(Intercept)          wt         cyl        disp          hp          am 
38.20279869 -3.30262301 -1.10637984  0.01225708 -0.02796002  1.55649163

User Input

Get the following input for a car:

wtIn: Weight of the car (in lbs/1000)
cylIn: No of Cylinders (4/6/8)
dispIn: Cylinder displacement volume (in Cu. In)
hpIn: Engine Gross horsepower (in Hp)
amIn: Types of transmission (0=Automatic / 1=Manual)

Predit Mileage

Create Dataframe with Input

df <- data.frame(wt=wtIn, cyl=cylIn, disp=dispIn, hp=hpIn, am=as.numeric(amIn))

Predicted Mielage

#Considering Default values..
df <- data.frame(wt=3.325, cyl=6, disp=196.3, hp=123, am=as.numeric(0))

predict(linearModel, newdata=df)

       1 
19.55028