predict price of car from vehicles dataset

Post on 14-Apr-2017

93 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

HIGH PERFORMANCE ANALYTICS USING

SAP HANAVehicle (dataset)

Project Team (Group 1):

- Sumit Kumar Saini

- Mohamed Salihdeen- Subhor Verma- Yogesh Dangi

DATA SET• There are two tables in our dataset• Table – 1 Vehicle Master

• It is our original vehicle dataset with all basic details about vehicle such as VIN, make, model, mileage, price, production date, no of services till date and age.

• Table – 2 Vehicle transactional data• It contains combined columns of other datasets to better visualize and

analyze the features and prediction. It contains primary column from master table i.e VIN and price. Other features are Insured amount, premium, service type, Dealer No, job class and max education.

HYPOTHESES• Hypothesis 1 – Does the age of the car and the mileage is making an

impact on the insurance amount of the car.

• Hypothesis 2 – Does production year or age of the car is impacting the premium for the car.

• Hypothesis 3- Does the model of the car affecting the premium amount.

• Hypothesis 4- Does the job class has any affect on the price of the car they purchase.

THEOREM 1

THEOREM 1 (CONTD.)

THEOREM 2

THEOREM 3

THEOREM 4

THEOREMS• Theorem I - We can see that as the age of the car is increasing, the

insurance amount of the car is decreasing.

• Theorem II - We can see that as the age of the car is increasing the premium for the car is decreasing.

• Theorem III - We can see that the expensive cars have higher premium. Like Audi has higher premium amount than Chevy cars

• Theorem IV – We can see that doctors or lawyers are purchasing more expensive cars than blue-collard people.

TOOLS USED

• SAP HANA - This is used for data extraction and building views ie:- attribute view and analytical views.

• SAP LUMIRA- This is used for building Visualizations.

• SAP Predictive Analytical tools – This is used for predictive analysis based on the input features.

SAP PREDICTIVE ANALYTICAL TOOL - MODELER

• We have regression analysis and decision tree to build the model and predict the values. Our target value is price.

SAP PREDICTIVE ANALYTICAL TOOL - MODELER

• Predictive power KI is 91.55% i.e model predict the price of car based on input features 91.55% accurately.

• Prediction confidence KR is 98.04% i.e model can be reliably used to continue on the data set.

SAP PREDICTIVE ANALYTICAL TOOL - MODELER

• Regression Model

SAP PREDICTIVE ANALYTICAL TOOL - MODELER

• Decision tree model

SAP PREDICTIVE ANALYTICAL TOOL - MODELER

• Statistical Report

SAP PREDICTIVE ANALYTICAL TOOL - MODELER

• Influence on target of input feature

SAP PREDICTIVE ANALYTICAL TOOL - MODELER

• Maximum Variable contribution

SAP PREDICTIVE ANALYTICAL TOOL - MODELER

• Model predicted values

SAP PREDICTIVE ANALYTICAL TOOL - MODELER

SAP PREDICTIVE ANALYTICAL TOOL – EXPERT ANALYTICS

• We have performed auto regression analysis in expert analytics keeping price as target variable.

SAP PREDICTIVE ANALYTICAL TOOL – EXPERT ANALYTICS

• Feature selection

SAP PREDICTIVE ANALYTICAL TOOL – EXPERT ANALYTICS

• Variable contribution

SAP PREDICTIVE ANALYTICAL TOOL – EXPERT ANALYTICS

• Model Graph

top related