1
$\begingroup$

Any input on my following query would be great.

Currently I have a a number of monthly datasets of individual(s) who have purchased properties (2012-2017). These individuals can be split into two categories. 1)First-time buyer 2)Non first_time buyer. Therefore, it is possible to create a monthly statistic on the percentage of home buyers who are first time buyers and who are not.

However, what in the future if this indicator was to disappear from my dataset, say it was no longer available in 2019. Would it be possible to use the previous data collected from 2012-2017 to predict who are first-time and non first time buyers into the future?

The dataset contains various information about these individuals and the property (age, gender, marital status, income, property location and property price).

I'm think of using a logistic regression to carry this out, but would be interested to hear of any papers which have carried something like this out before as I need to know what are good predictors to include in my model.

If anyone can think of any other issues please let me know.

$\endgroup$
1
  • $\begingroup$ So just to further this, would it be the case of developing a logistic regression model but with a decision rule built in at the end? For instance a first time buyer indicator but with a confidence intervals at 95%? – $\endgroup$
    – Sean_C
    Oct 4 '18 at 9:58
1
$\begingroup$

To me this sounds like a classification problem. There are a multitude of methods that can be used to classify variables into categories, and yes, logistic regression is one of them (given that you only have two categories that you want to classify something into).

As for recommendations of literature, I very much recommend The Elements of Statistical Learning by Hastie, Tibshirani and Friedman. It's a great book, well worth reading. It's also available for download from Hastie's homepage: https://web.stanford.edu/~hastie/ElemStatLearn/

References

Friedman, J., Hastie, T., & Tibshirani, R. (2011). The elements of statistical learning. Springer series in statistics. Springer, New York, 11th printing, 2nd edition.

$\endgroup$
2
  • 1
    $\begingroup$ Have a look at this: stats.stackexchange.com/questions/127042/… $\endgroup$ Oct 3 '18 at 11:23
  • $\begingroup$ So just to further this, would it be the case of developing a logistic regression model but with a decision rule built in at the end? For instance a first time buyer indicator but with a confidence intervals at 95%? $\endgroup$
    – Sean_C
    Oct 3 '18 at 14:02

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.