Please wait ...
0% Complete
Home
/
15th International Conference on Computer and Knowledge Engineering
PersianILP: Construction and Evaluation of a Standard Persian Dataset for Inductive Link Prediction
Authors :
Mohammad Rahimi
1
Afsaneh Fatemi
2
Ahmad Baraani
3
1- Dept. of Software Engineering
2- Dept. of Software Engineering
3- Dept. of Software Engineering
Keywords :
Inductive Link Prediction،Knowledge Graph Completion،PersianILP Dataset
Abstract :
Link prediction in knowledge graphs is a key task aimed at addressing the challenge of graph sparsity. In inductive link prediction, a model is trained on one graph and evaluated on another containing unseen entities. While twelve inductive datasets have been introduced for English to benchmark models in this domain, no such dataset exists for Persian. This study introduces PersianILP, the first Persian dataset designed for inductive link prediction. PersianILP is constructed through a purposeful combination of real-world data extracted from the FarsiBase knowledge graph and synthetic data generated using the DeepSeek language model. To evaluate PersianILP, key criteria such as structural and semantic diversity, statistical alignment between synthetic and real data, and adherence to inductive evaluation principles were considered. The dataset is compared with twelve benchmark datasets, including WN18RR, FB237, and NELL995. PersianILP contains 16,306 semantic triples, 10,693 entities, and 432 unique relations, exhibiting a highly sparse structure with a sparsity rate of 0.99. Evaluation using a baseline inductive link prediction model confirms the dataset’s high quality and effectiveness. Statistical analyses further demonstrate that PersianILP meets all essential requirements for research in inductive link prediction and can serve as a standard resource for studies in Persian language processing, semantic web, and recommender systems.
Papers List
List of archived papers
Optimizing Question-Answering Framework Through Integration of Text Summarization Model and Third-Generation Generative Pre-Trained Transformer
Ervin Gubin Moung - Toh Sin Tong - Maisarah Mohd Sufian - Valentino Liaw - Ali Farzamnia - Farashazillah Yahya
FarSick: A Persian Semantic Textual Similarity And Natural Language Inference Dataset
Zahra Ghasemi - Mohammad Ali Keyvanrad
Enhancing Lighter Neural Network Performance with Layer-wise Knowledge Distillation and Selective Pixel Attention
Siavash Zaravashan - Sajjad Torabi - Hesam Zaravashan
Hybrid navigation based on GPS data and SIFT-based place recognition using Biologically-inspired SLAM
Sahar Salimpour Kasebi - Hadi Seyedarabi - Javad Musevi Niya
Degarbayan-SC: A Colloquial Paraphrase Farsi Subtitles Dataset
Mohammad Javad Aghajani - Mohammad Ali Keyvanrad
Recommending Popular Locations Based on Collected Trajectories
Mohammad Rabbani bidgoli - Saber Ziaei
Enhanced Melanoma Detection: An Improved Deformable DETR Model with Efficient Channel Attention
Amirreza Rouhbakhshmeghrazi - Shayan Nalbandian - Sheida Shadman - Mohammad Reza Hassannezhad - Shuyuan Yang - Bo Li
Probabilistic Short-Term Load Forecasting Using GBDT-Based Sister Forecasts and Ensemble Methods
Hossein Shahinzadeh - Hamed Nafisi - Amirafshin Zamani - Saiedeh Mehrabani-Najafabadi - Arezou Mahmoudi - Farshad Ebrahimi
Graph Attention Networks for Modeling Multi-Sensor Relationships in Early Prediction of Critical Events in ICU Patients
Amir Akhavan Saffar - Danial Eskandari Faruji - Javad Hamidzadeh
A Framework for Automated Cardiovascular Magnetic Resonance Image Quality Scoring based on EuroCMR Registry Criteria
Shahabedin Nabavi - Mohsen Ebrahimi Moghaddam - Ahmad Ali Abin - Alejandro Frangi
more
Samin Hamayesh - Version 46.0.2