Urban Functional Area Identification via Stacking Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for identifying urban functional areas are inefficient and lack dynamic representation, failing to effectively capture social and economic information, and are subjective due to complex data structures and long data acquisition cycles.
Innovation Solution
An identification method based on the mixing degree of functions and integrated learning, which involves data preprocessing, constructing indicator features, training a Stacking-based integrated learning model using machine learning algorithms, and dividing the training dataset by mixing degree of functions to accurately map urban features to urban functional areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional geographic analysis methods are used to identify urban functional areas, then the identification process can be completed, but the processing efficiency is low and the method is subjective due to complex data structures
Solution Approach 1:
The patent replaces traditional geographic analysis methods with integrated learning algorithms including random forest, gradient boosting decision tree, support vector machine, and back-propagation neural network. This substitution of mechanical/geographic analysis with computational learning systems resolves the contradiction by achieving higher processing efficiency while objectively handling complex multi-source data structures.
Solution Approach 2:
The patent transforms the identification approach by changing from traditional geographic parameters to integrated learning model parameters. By using multiple algorithms with different parameter optimization capabilities, the system efficiently processes complex data structures and improves processing speed while maintaining objectivity.
2Loss of information
If remote sensing (RS) is used to acquire land use data, then physical changes in urban functional areas can be captured, but social and economic information cannot be presented
Solution Approach 1:
The patent merges remote sensing data with multiple other data sources including POI data, mobile signaling data, and online car-hailing demand data. This combination integrates physical information from RS with social and economic information from other sources, resolving the limitation of RS alone while maintaining a manageable system through unified data processing.
Solution Approach 2:
The patent creates a multi-functional data acquisition system that collects not only physical land use information but also social and economic data through multiple data sources. This universal approach allows the system to present both physical changes and socio-economic characteristics simultaneously.
3Productivity
If traditional identification methods are used, then the process can be completed, but the data acquisition cycle is long
Solution Approach 1:
The patent replaces traditional sequential identification methods with parallel integrated learning algorithms that can process multiple data sources simultaneously. This substitution dramatically reduces the identification cycle by enabling concurrent processing of remote sensing data, POI data, mobile signaling data, and other information sources.
Solution Approach 2:
The patent performs preliminary data preprocessing and feature extraction before the actual identification process. By preparing data in advance and organizing it into suitable formats for integrated learning, the system reduces the overall identification cycle time while maintaining high processing efficiency during the actual classification phase.
4Reliability
If traditional geographic analysis methods are used, then identification can be performed, but the method is subjective in threshold selection
Solution Approach 1:
The patent replaces subjective geographic analysis with objective integrated learning algorithms. The random forest, gradient boosting decision tree, support vector machine, and neural network models use mathematical optimization and statistical learning to determine thresholds automatically, eliminating human subjectivity while maintaining analytical rigor.
Solution Approach 2:
The integrated learning models incorporate feedback mechanisms where the algorithms learn from the data and continuously optimize their threshold selections. This feedback-driven approach ensures objective threshold determination based on actual data patterns rather than subjective judgment, improving reliability while managing complexity through automated learning.
Data Source
AI summary
An identification method of urban functional areas based on mixing degree of functions and integrated learning includes the following steps: 1) performing data acquisition and preprocessing; 2) constructing 10 indicator features of an urban functional area identification system; 3) structuring the indicator features: acquiring, by a spatial statistical tool, the 10 indicator features corresponding to each parcel; 4) constructing an independent variable dataset; 5) labeling response variables; 6) dividing a training dataset into a plurality of training subsets according to the mixing degree of functions; 7) training a Stacking-based integrated learning model; and 8) joining an attribute in one table to another table, so as to complete the identification of the urban functional areas on each parcel. The identification method divides the training dataset by grading the mixing degree of functions, and makes predictions based on the prediction dataset with corresponding mixing degree of functions.