Regularized Interval-valued Time Series Modeling for Crude Oil Price Forecasting

A penalized minimum distance estimation method for high-dimensional interval-valued time series is proposed. The method uses LASSO and adaptive LASSO techniques and is demonstrated via Monte Carlo simulations and empirical applications. The dataset, last updated on 2026-05-18, includes materials for forecasting crude oil prices and constructing sparse index-tracking portfolios.

Use Cases

• Forecasting crude oil prices based on interval-valued time series data mentioned in the description.
• Constructing sparse index-tracking portfolios based on the proposed regularized regression method.
• Benchmarking machine learning models like random forests and multilayer perceptrons for interval-valued data.
• Studying the consistency and oracle properties of penalized estimators in high-dimensional settings.

Strengths

• The methodology is validated against competing approaches including random forests and multilayer perceptrons.
• The proposed estimator's properties are established regardless of whether the number of predictors grows slower or faster than the sample size.
• The dataset is openly available under a CC-BY-4.0 license.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

figshare

Collection Method

The description suggests the data is generated via Monte Carlo simulations and empirical applications.

Freshness

Last updated 2026-05-18 15:42:08; freshness should be verified.