A Hybrid System for Temporal Relation Extraction from Discharge Summaries - Technologies and Applications of Artificial Intelligence - page 383

Tabl e 1. Performance of the rule-based and CRFs-based approaches on test set

Config. Precision Recall F-score
Rule-based 0.66 0.47 0.55

CRFs-based 0.59 0.63 0.61

Tabl e 2. Performance comparison of different temporal relation pair categories on
test set

Relation Category Precision Recall F-score
Event-sectime 0.88 0.73 0.80

Event-event 0.33 0.77 0.46

Event-time 0.62 0.75 0.67

Consecutive 0.68 0.46 0.55

Co-reference 0.32 0.42 0.36

5Con lu on

We developed a hybrid TLink extraction system combining a rule-based ap-
proach and a CRFs-based approach to extract temporal relations in clinical
records. In our rule-based approach, we define and apply a number of linguistic
rules. Our CRFs-based approach divides the TLink extraction tasks into three
sub-tasks and formulates them as sequence labeling problems. We use the above
linguistic rules as features to solve these problems. We evaluate our system on
the i2b2 2012 TLINK task dataset, and our results show that our approach
achieves a F-score of 61%.

References

1. Chang, Y.-C., Dai, H.-J., Wu, J.C.-Y., Chen, J.-M., Tsai, R.T.-H., Hsu, W.-L.:
TEMPTING system: A hybrid method of rule and machine learning for temporal
relation extraction in patient discharge summaries. Journal of Biomedical Informat-
ics 46(suppl.), S54-S60 (2013), 2012 i2b2 NLP Challenge on Temporal Relations in
Clinical Data
2. Kovaevi, A., Dehghan, A., Filannino, M., Keane, J.A., Nenadic, G.: Combining
rules and machine learning for extraction of temporal expressions and events from
clinical narratives. Journal of the American Medical Informatics Association (2013)
3. Lafferty, J.D., McCallum, A., Pereira, F.C.N.: Conditional random fields: Prob-
abilistic models for segmenting and labeling sequence data. In: Proceedings
of the Eighteenth International Conference on Machine Learning, ICML 2001,
pp. 282-289. Morgan Kaufmann Publishers Inc., San Francisco (2001)
4. Pustejovsky, J., Hanks, P., Sauri, R., See, A., Gaizauskas, R., Setzer, A., Radev,
D., Sundheim, B., Day, D., Ferro, L., Lazo, M.: The TIMEBANK corpus. In: Pro-
ceedings of Corpus Linguistics 2003, Lancaster, pp. 647-656 (March 2003)