Using Word Embeddings for Ontology Enrichment
Journal Title: International Journal of Intelligent Systems and Applications in Engineering - Year 2016, Vol 4, Issue 3
Abstract
Word embeddings, distributed word representations in a reduced linear space, show a lot of promise for accomplishing Natural Language Processing (NLP) tasks in an unsupervised manner. In this study, we investigate if the success of word2vec, a Neural Networks based word embeddings algorithm, can be replicated in an aggluginative language like Turkish. Turkish is more challenging than languages like English for complex NLP tasks because of her rich morphology. We picked ontology enrichment, again a relatively harder NLP task, as our test application. Firstly, we show how ontological relations can be extracted automaticaly from Turkish Wikipedia to construct a gold standard. Then by running experiments we show that the word vector representations produced by word2vec are useful to detect ontological relations encoded in Wikipedia. We propose a simple but yet effective weakly supervised ontology enrichment algorithm where for a given word a few know ontologically related concepts coupled with similarity scores computed via word2vec models can result in discovery of other related concepts. We argue how our algorithm can be improved and augmented to make it a viable component of an ontoloy learning and population framework.
Authors and Affiliations
İzzet Pembeci*| Muğla Sıtkı Koçman University. Department of Computer Engineering
Keywords
Related Articles
- EP ID EP799
- DOI 10.18201/ijisae.58806
- Views 395
- Downloads 25
A simple Mathematical Fuzzy Model of Brain Emotional Learning to Predict Kp Geomagnetic Index
In this paper, we propose fuzzy mathematical model of brain limbic system (LS) which is responsible for emotional stimuli. Here the proposed model is utilized to predict the chaotic activity of the earth’s magnetosphere....
Particle Swarm Optimization with Flexible Swarm for Unconstrained Optimization
Particle Swarm Optimization (PSO) algorithm inspired from behaviour of bird flocking and fish schooling. It is well-known algorithm which has been used in many areas successfully. However it sometimes suffers from premat...
New Approach in E-mail Based Text Steganography
In this study combination of lossless compression techniques and Vigenere cipher was used in text steganography that makes use of email addresses to be the keys to reconstruct the secret message which has been embedded i...
Rainfall Runoff Modelling Using Generalized Neural Network and Radial Basis Network
Rainfall runoff study has a wide scope in water resource management. To provide a reliable prediction model is of paramount importance. Runoff prediction is carried out using generalized regression neural network and rad...
A Mitigation Technique for Inrush Currents in Load Transformers for the Series Voltage Sag Compensator
In many countries, high-tech manufacturers concentrate in industry parks. Survey results suggest that 92% of interruption at industrial facilities is voltage sag related. An inrush mitigation technique is proposed and im...