good morning,

I am sure you are familiar with the titanic dataset predicting survival rate. I created regression model, one linear discriminant analysis model, one Naive Bayes model, and one K-Nearest Neighbor model to predict survival using the train set. I am suppose to predict the test set data using all four models. However, the test does not include the Survived variable. How can I complete this task?

Hi @Hesham Dabbas,

What is your source for the titanic.sav dataset?

I googled for it and found
https://github.com/datasciencedojo/datasets/blob/master/titanic.csv

The second variable in the file is "Survived", dichotomous, coded 0 and 1.   Is this the same dataset you are referencing?

Thank you. The instructor provided two separate files: train and test. The train include the Survived variable but the test does not. Please see attached. 

Attachment(s)

test.csv   29 KB 1 version

train.csv   56 KB 1 version

Original Message:
Sent: Thu May 23, 2024 11:07 AM
From: David Dwyer
Subject: Using train dataset to predict the test dataset

Hi @Hesham Dabbas,

What is your source for the titanic.sav dataset?

I googled for it and found
https://github.com/datasciencedojo/datasets/blob/master/titanic.csv

The second variable in the file is "Survived", dichotomous, coded 0 and 1.   Is this the same dataset you are referencing?

------------------------------
David Dwyer
SPSS Technical Support
IBM Software
------------------------------

Well, you can predict the outcomes in the test dataset (using the Scoring Wizard), and you can summarize those outcomes, but since you can't evaluate the quality of the predictions without the survived variables, maybe the instructor wants your predictions dataset.

You can compare the predictions across the different estimated models.  It would be interesting to see how well they agree.

Thank you. Do you have or recommend instruction on how to use the scoring wizard?