A Method for Identifying Japanese Shop and Company Names by Spatiotemporal Cleaning of Eccentrically Located Frequently Appearing Words.
Yuki AkiyamaRyosuke ShibasakiPublished in: Adv. Artif. Intell. (2012)
Keyphrases
- experimental evaluation
- high accuracy
- similarity measure
- pairwise
- synthetic data
- detection algorithm
- keywords
- preprocessing
- cost function
- dynamic programming
- knn
- detection method
- high precision
- case study
- missing data
- clustering method
- support vector machine svm
- input data
- mutual information
- co occurrence
- support vector machine
- computational cost
- significant improvement