| 摘要 |
PROBLEM TO BE SOLVED: To provide a more general-purpose and efficient initialization method for selecting a mixture model.SOLUTION: A method includes: determining a candidate model on the basis of a training sample set S110; generating a set of second initialized latent variables on the basis of one or more first initialized latent variables and the update latent variables of the candidate model S120, therein the candidate model is generated in accordance with the first initialized latent variables, and the update latent variables show the result of the grouping of samples output by the candidate model; and determining a target model on the basis of the set of the second initialized latent variables S130.SELECTED DRAWING: Figure 1 |
