Paper: arXiv 2506.15723

Authors: Irina G. Tanashkina, Alexey S. Tanashkin, Alexander S. Maksimchuik, Anna Yu. Poshivailo

Abstract

In this article, we review modern approaches to building interpretable models of property markets using machine learning on the base of mass valuation of property in the Primorye region, Russia. The researcher, lacking expertise in this topic, encounters numerous difficulties in the effort to build a good model. The main source of this is the huge difference between noisy real market data and ideal data which is very common in all types of tutorials on machine learning. This paper covers all stages of modeling: the collection of initial data, identification of outliers, the search and analysis of patterns in the data, the formation and final choice of price factors, the building of the model, and the evaluation of its efficiency. For each stage, we highlight potential issues and describe sound methods for overcoming emerging difficulties on actual examples. We show that the combination of classical linear regression with interpolation methods of geostatistics allows to build an effective model for land parcels. For flats, when many objects are attributed to one spatial point the application of geostatistical methods is difficult. Therefore we suggest linear regression with automatic generation and selection of additional rules on the base of decision trees, so called the RuleFit method. Thus we show, that despite such a strong restriction as the requirement of interpretability which is important in practical aspects, for example, legal matters, it is still possible to build effective models of real property markets.

Complexity vs Empirical Score

  • Math Complexity: 6.0/10
  • Empirical Rigor: 7.5/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper uses advanced statistical methods (regression-kriging, RuleFit) for spatial modeling and mentions graph theory, indicating moderate-to-high math complexity. It is heavily data/implementation-focused, covering full real-world data pipeline from collection to outlier detection and model evaluation on specific property types (flats/land parcels), demonstrating high empirical rigor.

Research Flowchart

  flowchart TD
  A["Research Goal: Build interpretable ML models<br>for mass property valuation in Primorye"] --> B{"Data Processing &<br>Feature Engineering"}
  B --> C["Geostatistical Interpolation<br>(for Land Parcels)"]
  B --> D["RuleFit (Linear Regression +<br>Decision Tree Rules)<br>(for Flats)"]
  C --> E["Computational Process<br>Model Building & Validation"]
  D --> E
  E --> F["Key Findings/Outcomes<br>Effective interpretable models achieved despite data noise.<br>Land: Linear regression + Geostatistics.<br>Flats: RuleFit allows interpretability with complex data."]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes using linear regression combined with kriging for land parcels and a RuleFit method (linear regression plus decision‑tree rules) for flats to build interpretable property‑price models.

Source sentence
We show that the combination of classical linear regression with kriging (interpolation method of geostatistics) allows to build an effective model for land parcels.

Assumes 2

  • The data must satisfy spatial continuity and stationarity for kriging to be applicable. statistical
    Source sentence
    There are two basic conditions of usage of these methods – the data should satisfy the spatial continuity and the stationarity conditions (Kovalevskiy, 2012).
  • Residuals are approximately normally distributed, allowing stationarity to be assumed via histogram inspection. statistical
    Source sentence
    If the histogram is close to the normal distribution, then the stationarity (statistical homogeneity) is assumed.

Data

  • Mass cadastral valuation of property in the Primorye region other unknown
    700 000 flats and more than 570 000 land parcels
    Source sentence
    The models presented in this work were used for cadastral valuation of 700 000 flats and more than 570…

Costs

Not applicable The paper reports no trading or execution result.

Finds 3

  • Regression‑kriging improves prediction accuracy for land parcels compared with a plain linear model. predictive-accuracy
    Source sentence
    We show that the implementation of this method for modeling of land parcels leads to improvement of predictions comparing to the linear model.
  • RuleFit achieves a coefficient of determination of 0.6 and lower MAE and MAPE than linear regression for flat price prediction. performance
    Source sentence
    The coefficient of determination is equal to 0.6, which is acceptable for our purposes. The values of MAE and MAPE for the RuleFit are lower than for the linear regression.
  • Interpretable models (regression‑kriging and RuleFit) perform similarly to the black‑box Random Forest method. performance
    Source sentence
    We compare the performance of our inherently interpretable models with well-proven "black-box" Random Forest method and demonstrate similar results.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 25 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.