Aim of the project
Support the retail real estate portfolio revaluation process with a new, innovative demand based and data mining based methodology.
Methodology
With advanced clustering algorithm (DBSCAN and Kohonen), we can separate the country into smaller, homogenous living areas. With decision trees, we can create real estate sub-segments regarding to its type, its size, etc. With these living areas, subsegments and collected advertisement data, we use regression and hedonic regression model to achieve our goal: the statistical based revaluation modeling.
Results
Revaluation of real estate portfolio with statistical methodology. It can be use to validate the existing methodology or create a new methodology.
DOs and DON’Ts
Do
- Ensure the cooperation of appraisers
- Good data quality is recommended to ensure the best results
|
|
Don’t
- Some data is must (e.g. basis area, plot area, location, etc.), some is welcomed (e.g. condition of the property, type of heating, etc.). If you do not satisfy the basic data needs, the project can not be done
- We don’t recommend statistical approach, if you have portfolio with very special real estate (e.g. luxury houses, etc.)
|
Why we are different
- We do not insist on the territorial hierarchy of Hungary.
- We create a more accurate territorial clustering according to data density, type of real estate and geo-location: the living areas.
|
