Abstract:Interval-valued data receives much attention due to its wide applications in the fields of finance, econometrics, meteorology and medicine. However, most regression models developed for interval-valued data assume observations are mutually independent, not adapted to the scenario that individuals are spatially correlated. We propose a new linear model to accommodate to areal-type spatial dependency existed in interval-valued data. Specifically, spatial correlation among centers of responses are considered. To improve the new model's prediction accuracy, we add three inequality constrains. Parameters are obtained by an algorithm combining grid search technique and the constrained least squares method. Numerical experiments are designed to examine prediction performances of the proposed model. We also employ a weather dataset to demonstrate usefulness of our model.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is that most of the existing regression models for processing interval - valued data assume that the observations are independent of each other, without considering the possible spatial correlation of data in practical applications. For example, in meteorological data, the seasonal precipitation in a city is often related to the precipitation in neighboring cities; in unemployment rate data, the unemployment rates in different areas of the same city also show spatial clustering. The data in these problems have obvious spatial dependence, that is, the smaller the distance between two spatial units, the stronger the spatial dependence between them. Therefore, this paper proposes a new linear model - the Constrained Spatial Autoregressive Model (ICSM) to adapt to the regional - type spatial dependence existing in interval - valued data. Specifically, this model mainly solves the following points: 1. **Spatial correlation**: It takes into account the spatial correlation between the central values of the response intervals and models this correlation by introducing the Spatial Autoregressive Model (SAR). 2. **Prediction accuracy**: In order to improve the prediction accuracy of the model, three inequality constraint conditions are added. These constraint conditions ensure that the predicted interval has an overlapping area with the true value and that the predicted radius value is positive. 3. **Parameter estimation**: An algorithm combining grid - search technology and constrained least - squares method is proposed to estimate the model parameters to adapt to the situation where the parameters are subject to inequality constraints. Through these improvements, the ICSM model can better handle interval - valued data with spatial correlation, thus providing more accurate prediction results in practical applications. The paper also verifies the effectiveness and superiority of this model through numerical experiments and the application of actual weather data.

A Constrained Spatial Autoregressive Model for Interval-valued data

Spatial Autoregressive Model for Compositional Data

A spatial Durbin model for interval-valued data with t-distribution

A Vector Autoregressive Moving Average Model for Interval-Valued Time Series Data

A Constrained Interval-Valued Linear Regression Model: A New Heteroscedasticity Estimation Method

Threshold Autoregressive Models for Interval-Valued Time Series Data

Nonparametric Estimation and Forecasting of Interval-Valued Time Series Regression Models with Constraints

Variable Selection of Spatial Logistic Autoregressive Model with Linear Constraints

Autoregressive Conditional Models for Interval-Valued Time Series Data

Functional Spatial Autoregressive Models

Valid model-free spatial prediction

Sparse Interval-valued Time Series Modeling with Machine Learning

Prediction for spatio-temporal models with autoregression in errors

Autoregressive Model with Spatial Dependence and Missing Data

A Robust Spatial Autoregressive Scalar-on-function Regression with T-Distribution.

Vector Autoregressive Models with Spatially Structured Coefficients for Time Series on a Spatial Grid

Local Walsh-average Regression for Spatial Autoregression Single Index Varying Coefficient Models

Estimation and testing of a higher-order partially linear spatial autoregressive model

Estimation of a Semiparametric Varying-Coefficient Mixed Regressive Spatial Autoregressive Model

Local linear regression for non grid spatiotemporal models with autoregressive errors

A spatial autoregressive model with a nonlinear transformation of the dependent variable