NTCIR-10 Workshop SpokenDoc-2 Task Formal Run Evaluation Results

[NTCIR Home]

* Evaluation Results

• Evaluation results for the Spoken Term Detection (STD) task
• For the moderate-size task ( eight groups with 27 runs + three baseline runs)
• Evaluation results for each participating group and the baseline including result's overview [SpokenDoc2-STD-SDPWS.xlsx]
• For the large-size task ( five groups with 21 runs + three baseline runs )
• Evaluation results for each participating group and the baseline including result's overview [SpokenDoc2-STD-CSJ.xlsx]
• Evaluation results for the inexistent Spoken Term Detection (iSTD) task
• Four groups with 18 runs and the three baseline runs were submitted.
• Evaluation results for each participating group and the baseline including result's overview [SpokenDoc2-iSTD.xlsx]
• Evaluation results for the Spoken Content Retrieval (SCR) subtask
• Lecture Retrieval Task ( six groups with 34 runs + 4 baseline runs )
• Evaluation results for each participating group and the baseline [SpokenDoc2-SCR-LEC.xlsx]
• Passage Retrieval Task ( five groups with 35 runs + baseline runs )
• Evaluation results for each participating group and the baseline [SpokenDoc2-SCR-PAS.xlsx]

 

These results were used in the overview paper:

T. Akiba, H. Nishizaki, K. Aikawa, X. Hu, Y. Itoh, T. Kawahara, S. Nakagawa, H. Nanjo, and Y. Yamanashita (2013) Overview of the NTCIR-10 SpokenDoc-2 task. In Proceedings of the 10 ^th NTCIR Conference on Evaluation of Information Access Technologies, Tokyo, June 18-21, 2013.