The car and different specs

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Problem statement.

As we are studying the evaluation of the car and different specs of the automobiles. All of our data based on the different specifications of different cars.

But here we face another difficulty that if we choose one car manufacturer and choose a specific model of the car and gather the specification s of that model like its condition year of registration and its doors etc. its not necessary that all of the other model of same specification from the same manufacturer will perform in a same way.

For example if we choose a specific size of engine power in different models of a car from a single manufacturer. The performances of these different models with equal engine size will be different.

Its quite strange if we think that all of these cars are from same manufacturer and have same engine powers and showing different behaviours.

Because its not only one spec of the specific car which we are going to consider. For that reason we have to take a deep look on the automobile to configure some more basic characteristics which can play an important role in the information which we are going to gather for our datasets.

Our decision will be based on our datasets which we will gather on the basis of the characteristics.

The model evaluate cars in following fashion.

1. Car's general characteristics :

  • Car acceptability.
  • Overall buying price.
  • Price of the maintenance.

2. Technical characteristics :

  • Comfort.
  • Number of doors.
  • Person capacity.
  • Boot size.
  • Level of safety equipments.

As we got two main categories of the characteristics of the automobiles so we can create our datasets by using theses specifications.

Real world problem:

Now here we face the real world problem that how can we find out the datasets credibility by taking any specific model.

For that reason we took six main attributes picking key specs from both general and technical characteristics such as number of doors, safety equipments, boot size, buying condition, person capacity, acceptable condition.

Now we use J48 algorithm to resolve this problem.

Implementation of data set

  • attributes
  • buying
  • maint
  • doors
  • persons
  • lug_boot
  • safety
  • car

These attributes are defined in an ARFF file, the relation name is auto_mpg. File is defined as follows:

  • @relation auto_mpg
  • @attribute buying {vhigh,high,med,low}
  • @attribute maint {vhigh,high,med,low}
  • @attribute doors {2,3,4,5more}
  • @attribute persons {2,4,more,}
  • @attribute lug_boot {small,med,big}
  • @attribute safety {low,med,high}
  • @attribute car {unacc,acc,vgood,good}

Instances:

Implementing Classifier:

Now we are going to create decision tree by using one of the algorithm which are available in weka. So we are going to create it in J48 algorithm.

This is available in weka and it's another version of C45. In Weka's classifier package J48 is available.[4]

C45 and J48 algorithms generates decision tree which help us out to sort out our data.

Decision Tree.

Decision tree is very popular tool which is mainly popular for data classification and making predictions about the data which is presented to the problem.

Decision tree gives the procedures and rules which make things easy for the person for given problem to sort it out.

How to do J48.

As we define earlier we create decision tree by using J48 algorithm.

These are the steps to perform it.

  • Open an ARFF file in Weka as we normally open any file.
  • Select Weka the Classify tab.
  • Make the selection of it.
  • Select trees for J48.
  • Click start button.
  • Select and right click on the first element on the which appear under the start button.
  • Now to show the tree so click on visualize tree.
  • To make it fit in screen select fit to screen button.

By using this data and the ARFF file we got. We make decision tree. So a tree look like this will be created.

Conclusion.

While we use decision tree to generate any results of the given data for the problem it is normally very efficient in nature.

But it is not necessary to over come the problem properly. But this is the only case when it comes on real world problem and real data. But its not necessary that it works like this every time.

For our problem here which we applied this algorithm its show the result that if the cars condition is not acceptable then the percentage of its non sellable will be very high .

But if it is in proper acceptable condition then the percentage of its non selling will be very low.

So in this particular case it works precisely and efficiently.

References.

  1. Weka 3 - Data Mining Tool with Open Source Machine Learning Software in Java [Online] available at http://www.cs.waikato.ac.nz/ml/weka/ [Accessed on 20 Nov 2009]
  2. My Weka Page [Online] available at http://www.hakank.org/weka/ [Accessed on 21 Nov 2009]
  3. Attribute Relation File Format (ARFF) [Online] available at http://www.cs.waikato.ac.nz/~ml/weka/arff.html [Accessed on 23 Nov 2009]
  4. Classification via Decision Trees in Weka [Online] available at http://maya.cs.depaul.edu/~classes/ect584/WEKA/classify.html [Accessed on 30 Nov 2009]
  5. Teachers slides and material provided in class