Outlier Season is Open, Bring Your Chainsaw
Presentation Category
Modeling and Valuation
General Session Description
A common struggle in mass appraisal is dealing with data that contains outliers and incorrect information. When trying to remove outliers we wish to do so in a way that makes the data truly more accurately represent the population, while doing this it can be hard to truly determine what is an outlier and what is not without adding bias. Isolation Forests offer the ability for us to use an unsupervised machine learning technique to identify outliers in our data. In this presentation we will discuss how Isolation Forests work, how they can be applied to Mass Appraisal and some actual use cases of Isolation Forests in mass appraisal.
Start Date
4-1-2026 10:30 AM
End Date
4-1-2026 12:00 PM
Outlier Season is Open, Bring Your Chainsaw
A common struggle in mass appraisal is dealing with data that contains outliers and incorrect information. When trying to remove outliers we wish to do so in a way that makes the data truly more accurately represent the population, while doing this it can be hard to truly determine what is an outlier and what is not without adding bias. Isolation Forests offer the ability for us to use an unsupervised machine learning technique to identify outliers in our data. In this presentation we will discuss how Isolation Forests work, how they can be applied to Mass Appraisal and some actual use cases of Isolation Forests in mass appraisal.