| 摘要 |
A system is configured to receive a word that includes a group of samples; randomly select a subset of the samples; identify first samples, from the subset, with a lowest level of reliability; select another subset of the samples; identify second samples, from the other subset, with a lowest level of reliability; and create a merged subset based on selected first samples and selected second samples. The system is also configured to select a further subset of the samples; identify third samples, from the further subset, with a lowest level of reliability; identify fourth samples, from the merged subset, associated with a lowest level of reliability; create another merged subset based on a greater probability that fourth samples than third samples are included in the other merged subset; and generate another word based a sample from the other merged subset; and process the word using the other word.
|
