POI Name Identification Using Segmented Machine Learning Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for identifying names of Points of Interest (POI) in navigation systems suffer from low accuracy due to the dynamic nature of specific expressions in various fields, making it challenging to extract information effectively.
Innovation Solution
A search device and method utilizing a POI presence/absence learning model, a POI state learning model, and a POI name learning model, which use word feature vector models to determine the presence, state, and name of POI with higher accuracy by processing document groups through morpheme analysis and machine learning algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional methods are used to identify POI names from posted data, then operator effort is reduced, but identification accuracy is insufficient
Solution Approach 1:
The identification process is divided into three sequential stages: (1) POI presence/absence determination using a first determination unit, (2) POI state determination using a second determination unit, and (3) POI name identification using an identifying unit. This segmentation allows each stage to focus on specific aspects, improving overall accuracy while maintaining automation.
Solution Approach 2:
The system performs preliminary filtering by determining POI presence/absence and state before final name identification. The POI presence/absence learning model and POI state learning model pre-process the data to identify relevant documents, ensuring that the name identification stage only processes high-probability candidates, thereby improving accuracy.
2Adaptability or versatility
If specific expressions change day by day in various fields, then adaptability is improved, but extraction of information becomes unrealistic without generating dictionaries
Solution Approach 1:
The system uses machine learning models with learnable parameters that automatically adapt to changing expressions. The POI presence/absence learning model, POI state learning model, and POI name learning model are trained on data and continuously improve their ability to recognize new expressions without requiring manual dictionary updates. This parameter-based adaptation eliminates the need for complex dictionary generation while maintaining high adaptability.
Solution Approach 2:
The machine learning models perform self-learning and self-improvement by processing training data and adjusting their internal parameters automatically. The system extracts features from documents, determines POI presence and state, and identifies names without human intervention in the extraction process, making the system self-sufficient and eliminating the need for manual dictionary creation.
3Measurement precision
If machine learning models are used to determine POI presence, state, and name, then identification accuracy is improved, but processing complexity increases
Solution Approach 1:
The complex machine learning process is segmented into three separate models: POI presence/absence learning model, POI state learning model, and POI name learning model. Each model handles a specific aspect of the identification process, reducing the complexity of individual models while maintaining high overall accuracy through their coordinated operation.
Solution Approach 2:
The system extracts only the necessary features from documents for each determination stage. The POI presence/absence determination extracts features relevant to identifying whether a POI is mentioned, the state determination extracts features about the POI's condition, and the name identification extracts the actual POI name. This selective extraction reduces processing complexity while maintaining accuracy.
Data Source
AI summary
A search device identifies names of POI from a document group having not been determined. A storage unit that stores a POI presence/absence learning model having learned contexts relating to presence/absence of POI, a POI state learning model having learned contexts relating to states of POI, and a POI name learning model having learned features relating to names of POI, an acceptance unit that accepts a first document group that is a determination target, first and second determination units and an identifying unit that identifies a name of a POI using the POI name learning model from each document of a third document group for which information relating to states of POI is determined by the second determination unit in a second document group are included.