Tabular Deep Learning Models Compared to Classical ML for Land Cover Classification

A new research paper compares the effectiveness of tabular deep learning (TDL) models against classical machine learning algorithms for urban land cover classification. The study utilized the ULC dataset from the UCI Machine Learning Repository, which contains tabular features derived from aerial imagery. Researchers benchmarked models like logistic regression, support vector machines, random forests, XGBoost, and Catboost against TDL models including TabNet, FT-Transformer, TabTransformer, TabSeq, and 1D CNNs. Results indicated that while tree ensembles performed well, TDL models could achieve comparable or superior performance, particularly when dealing with significant non-linear interactions and class imbalance. AI

IMPACT This research provides insights into the performance of deep learning models on tabular data for land cover classification, potentially informing future applications in urban planning and environmental monitoring.

RANK_REASON Research paper comparing machine learning models for a specific task. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• 1D CNNs
• Al Zadid Sultan Bin Habib
• Catboost
• FT-Transformer
• logistic regression model
• random forest
• support vector machine
• TabNet: Attentive Interpretable Tabular Learning
• TabSeq
• TabTransformer: Tabular Data Modeling Using Contextual Embeddings
• UCI Machine Learning Repository
• XGBoost

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →