Data Science Dojo
Copyright (c) 2016 - 2019
Level: Beginner
Recommended Use: Classification Models
Domain: Automobile
Car Evaluation Data Set
Predict acceptibility of a car
This intermediate level data set was derived from a decision-making model which was originally developed for research on multi-attribute decision making. Decision making involves selection between seemingly conflicting alternatives.
The data set has 1728 rows and 7 columns in which car attributes such as price and technology are described across 6 attributes such as "Buying Price", "Maintenance", and "Safety" etc. There are multiple alternatives under each of the 6 attributes. Car's acceptability, the seventh attribute, is the outcome variable.
This data set is recommended for learning and practicing your skills in classification modelling techniques. Feel free to explore the data set with multiple classification methods. The Following data dictionary gives more details on this data set:
Data Dictionary
Column Position | Atrribute Name | Definition | Data Type | Example | % Null Ratios |
---|---|---|---|---|---|
1 | buying | Buying price of the car (v-high, high, med, low) | Qualitative | low, med, high | 0 |
2 | maint | Price of the maintenance of car (v-high, high, med, low) | Qualitative | low, med, high | 0 |
3 | doors | Number of doors (2, 3, 4, 5-more) | Qualitative | 2, 3, 4 | 0 |
4 | persons | Capacity in terms of persons to carry (2, 4, more) | Qualitative | 2, 4, more | 0 |
5 | lug_boot | The size of luggage boot (small, med, big) | Qualitative | small, med, big | 0 |
6 | safety | Estimated safety of the car (low, med, high) | Qualitative | low, med, high | 0 |
7 | class | Car acceptability (unacc: unacceptible, acc: acceptible, good: good, v-good: very good) | Qualitative | unacc, acc, good | 0 |
Acknowledgement
This data set has been sourced from the Machine Learning Repository of University of California, Irvine Car Evaluation Data Set (UC Irvine). The UCI page mentions following as the donor of the dataset:
- Marko Bohanec (marko.bohanec '@' ijs.si)
- Blaz Zupan (blaz.zupan '@' ijs.si)