% ==========================================================================
\section{Introduction}
\label{sec:introduction}
% ==========================================================================

The theory of hedonic prices, formalized by \citet{Rosen1974} and rooted in
the characteristics approach to consumer demand developed by
\citet{Lancaster1966}, posits that differentiated goods can be described by a
vector of objectively measured attributes and that the market price of such a
good equals the sum of the implicit valuations consumers place on each
attribute. In the four decades since Rosen's seminal contribution, hedonic
models have become a workhorse of applied microeconomics, with extensive
applications in real estate \citep{ChattopadhyayMachemehl2006}, automobiles,
environmental valuation, and, more recently, the sharing economy.

The rapid global expansion of Airbnb and competing short-term rental (STR)
platforms has created a uniquely rich laboratory for hedonic analysis. Unlike
traditional hotel room inventories, which are relatively standardized within
a star-rating tier, Airbnb listings exhibit enormous heterogeneity along
structural, locational, amenity, and reputational dimensions. Each listing
is, in effect, a bundle of characteristics chosen and priced by an individual
host operating in a decentralized marketplace. The platform's publicly
accessible data---including detailed descriptions, amenity lists, review
scores, host profiles, calendar availability, and geographic
coordinates---provides researchers with an unusually granular view of the
hedonic price surface.

A growing empirical literature has exploited this richness.
\citet{WangNicolau2017} estimate hedonic price equations for a cross-section
of Airbnb listings in 33~major cities, finding that property type, location,
host characteristics, and online reputation are significant price
determinants. \citet{Gibbs2018} extend the analysis to include amenity
bundles and find that premium amenities such as pools and hot tubs
command substantial price premiums. \citet{Zhang2017} and \citet{Chen2020}
document similar patterns across Chinese and U.S.\ markets, respectively,
while \citet{BenitezAurioles2018} and \citet{Dogru2017} examine European and
Turkish markets. \citet{PerezSanchez2018} apply geographically weighted
regression to Spanish Airbnb data, demonstrating that implicit prices vary
spatially. \citet{GunterOnder2018} study Vienna's market structure,
while \citet{Magno2018} investigate the role of host experience in shaping
pricing outcomes.

Despite these contributions, several gaps remain. First, most studies rely on
a single econometric specification---typically ordinary least squares (OLS)
applied to a log-linear price equation---and therefore cannot capture
heterogeneity across the conditional price distribution. Second, the
potential for complex non-linear interactions among listing characteristics
is seldom explored, and when machine learning methods are employed, the
resulting models are treated as black boxes rather than as complements to
structural hedonic inference. Third, spatial dependence in listing
prices---a first-order concern in any market where location matters---is
frequently acknowledged but rarely modeled explicitly using spatial
econometric techniques.

This paper addresses all three gaps by developing a comprehensive
multi-method hedonic pricing model for Airbnb listings in Quebec, Canada.
Our central research questions are:
%
\begin{enumerate}[leftmargin=2em]
    \item What are the implicit prices of structural, amenity, host,
          reputation, locational, and text-derived characteristics of
          Airbnb listings in Quebec?
    \item How do these implicit prices vary across the conditional price
          distribution, from budget to luxury segments?
    \item To what extent do non-linear interactions and spatial dependence
          alter the hedonic price surface estimated by conventional linear
          models?
    \item Which listing attributes contribute most to predictive accuracy,
          and does the importance ranking differ between econometric and
          machine learning frameworks?
\end{enumerate}

Our contribution is fourfold. First, we assemble a comprehensive hedonic
dataset for Quebec---a market that, despite its significance as a major
North American tourism destination, has received relatively little attention
in the Airbnb pricing literature. Our sample encompasses 5,000 listings
across Montr\'{e}al, Quebec City, the Laurentians, the Eastern Townships,
Charlevoix, and dozens of smaller municipalities. Second, we construct a
feature space spanning seven attribute categories---structural
characteristics, amenity indices, host attributes, reputation signals,
locational variables, NLP-derived text features, and spatial
controls---producing 49 modeling features. Third, we implement a layered
estimation strategy combining OLS with heteroskedasticity-robust standard
errors (HC3), quantile regression at five percentiles, an ensemble of seven
machine learning models interpreted through SHAP values
\citep{Lundberg2017}, and spatial econometric models including spatial lag,
spatial error, and spatial Durbin specifications \citep{LeSagePace2009,
Anselin1988}. Fourth, we demonstrate how estimated implicit prices aggregate
into a price decomposition that benchmarks any listing against the market,
providing actionable intelligence for hosts and platform managers.

The remainder of the paper is organized as follows.
Section~\ref{sec:data} describes the data sources, variable construction,
and summary statistics. Section~\ref{sec:methodology} presents the
econometric and machine learning methodology.
Section~\ref{sec:results} reports the main empirical results.
Section~\ref{sec:robustness} discusses robustness checks.
Section~\ref{sec:conclusion} concludes.