This site uses different types of cookies, including analytics and functional cookies (its own and from other sites). To change your cookie settings or find out more, click here. If you continue browsing our website, you accept these cookies.
I am currently in the process of building a few predictive models to help answer the following questions:
- Which current season ticket holders will renew?
- Who will buy a hospitality product?
- Who will buy a match ticket/season ticket?
For the latter two questions, I am suffering from rare event bias. This is because the number of e.g. hospitality purchasers relative to non-purchasers is tiny; around 500 purchasers relative to a population of 1M (average propensity: 0.0005). Using a standard probability threshold of 0.5 leads to a confusion matrix failing to predict any actual historical purchasers.