1
$\begingroup$

I have dataset which describe "how many passenger arriving in some airport " and I would like to predict how many passenger arriving in monthly bases for next year. The features that I have is the following :

year, month, airport, number of passenger (monthly)

In the data, three out of of 50 airports are usually have huge passengers arriving.

I have used random forest classifier but the issue that i'm encounter is I have RMSE is high. As a result i see huge different between the actual value and predicting value. How to fix this issue ?

$\endgroup$

2 Answers 2

1
$\begingroup$

The problem is that you are using a classifier. You should use RandomForestRegressor or other type regressors.

Those three airports seems like leverage points. Check this wiki topic. You can log transform your target values or simply remove those 3 airports from your training set.

But still you have very less observations. You may want to add new data to your set. After train test split 35 observations for train 15 for test seems not so satisfactory.

$\endgroup$
0
$\begingroup$

You probably need to re-design the problem in a way which takes time into account (time series), and maybe use a specific model for that.

Currently each of your instances is for a single month, so the algorithm tries to predict the number of passengers based only on this individual month and year. Logically the main information that the system would need is the number of passengers in the past few months for this airport, and it doesn't have it.

A simple way to start would be to format the data so that one instance contains features providing information about the last N months.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.