Limits...
Harnessing ontology and machine learning for RSO classification

View Article: PubMed Central - PubMed

ABSTRACT

Classification is an important part of resident space objects (RSOs) identification, which is a main focus of space situational awareness. Owing to the absence of some features caused by the limited and uncertain observations, RSO classification remains a difficult task. In this paper, an ontology for RSO classification named OntoStar is built upon domain knowledge and machine learning rules. Then data describing RSO are represented by OntoStar. A demo shows how an RSO is classified based on OntoStar. It is also shown in the demo that traceable and comprehensible reasons for the classification can be given, hence the classification can be checked and validated. Experiments on WEKA show that ontology-based classification gains a relatively high accuracy and precision for classifying RSOs. When classifying RSOs with imperfect data, ontology-based classification keeps its performances, showing evident advantages over classical machine learning classifiers who either have increases of 5 % at least in FP rate or have decreases of 5 % at least in indexes such as accuracy, precision and recall.

No MeSH data available.


Related in: MedlinePlus

Finding justifications for a classification
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC5037103&req=5

Fig4: Finding justifications for a classification

Mentions: Given an unlabelled RSO with data description, it can be classified via R-OBC and find a classification justification, by the following steps. Firstly, the RSO is represented in OntoStar as described in “Demo on Protégé of R-OBC” section. Then the ontology reasoning service realization, provided by the ontology reasoner pellet, is employed to compute the most specific class for the RSO. Finally, the service of explanation is employed to find a classification justification. An example of classifying an RSO named “inst2cls” (identified as inst2cls in the following) through R-OBC and finding a classification justification is shown in Fig. 4. The reasoning service realization is completed within 1840 ms.Fig. 4


Harnessing ontology and machine learning for RSO classification
Finding justifications for a classification
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC5037103&req=5

Fig4: Finding justifications for a classification
Mentions: Given an unlabelled RSO with data description, it can be classified via R-OBC and find a classification justification, by the following steps. Firstly, the RSO is represented in OntoStar as described in “Demo on Protégé of R-OBC” section. Then the ontology reasoning service realization, provided by the ontology reasoner pellet, is employed to compute the most specific class for the RSO. Finally, the service of explanation is employed to find a classification justification. An example of classifying an RSO named “inst2cls” (identified as inst2cls in the following) through R-OBC and finding a classification justification is shown in Fig. 4. The reasoning service realization is completed within 1840 ms.Fig. 4

View Article: PubMed Central - PubMed

ABSTRACT

Classification is an important part of resident space objects (RSOs) identification, which is a main focus of space situational awareness. Owing to the absence of some features caused by the limited and uncertain observations, RSO classification remains a difficult task. In this paper, an ontology for RSO classification named OntoStar is built upon domain knowledge and machine learning rules. Then data describing RSO are represented by OntoStar. A demo shows how an RSO is classified based on OntoStar. It is also shown in the demo that traceable and comprehensible reasons for the classification can be given, hence the classification can be checked and validated. Experiments on WEKA show that ontology-based classification gains a relatively high accuracy and precision for classifying RSOs. When classifying RSOs with imperfect data, ontology-based classification keeps its performances, showing evident advantages over classical machine learning classifiers who either have increases of 5 % at least in FP rate or have decreases of 5 % at least in indexes such as accuracy, precision and recall.

No MeSH data available.


Related in: MedlinePlus