Method and apparatus for expansion of search queries on large vocabulary continuous speech recognition transcripts

1. A method for expansion of search queries on large vocabulary continuous speech recognition transcripts comprising:

obtaining a textual transcript of audio interaction generated by the large vocabulary continuous speech recognition;

generating a topic model from the textual transcripts; said topic model comprises a plurality of topics wherein each topic of the plurality of topics comprises a list of keywords;

obtaining a search term;

associating a topic from the topic model with the search term; and

generating a list of candidate term expansion words by selecting keywords from the list of keywords of the associated topic; said candidate term expansion words are of high probability to be substitution errors of the search term that are generated by the large vocabulary continuous speech recognition.

2. The method according to claim 1 further comprises extracting a phonetic transcription of the search term.

3. The method according to claim 1 further comprises detecting the frequency of the search term in the textual transcripts.

4. The method according to claim 1 further comprises extracting phonetic transcriptions of the keywords on the list of keywords.

5. The method according to claim 1 further comprises extracting the stem form of the keywords on the list of keywords.

6. The method according to claim 1 further comprises detecting the frequency of the keywords on the list of keywords by counting the number of appearances of the keywords in the textual transcripts.

7. The method according to claim 1 wherein associating a topic from the topic model with the search term is performed by detecting keywords on the lists of keywords that are similar to the search term.

8. The method according to claim 7 wherein more than one keyword is detected along with a frequency of the detected keywords within each topic and wherein the search term is associated with the topic that comprises the detected keyword with the highest frequency.

9. The method according to claim 7 wherein more than one keyword is detected along with a weight of the detected keywords within each topic and wherein the search term is associated with the topic that comprises the detected keyword with the highest weight.

10. The method according to claim 1 wherein generating the list of candidate term expansion words is based on detecting phonetic similarities between the search term and the keywords within the associated topic.

11. The method according to claim 10 wherein the detection of the phonetic similarities comprises:

comparing between the phonetic transcription of the search term and the phonetic transcriptions of the keywords within the associated topic;

producing similarity scores; and

comparing between the similarity scores and a predefined similarity threshold.

12. The method according to claim 11 further comprises generating a list of term expansion words by selecting words from the list of candidate term expansion words.

13. The method according to claim 12 wherein selecting the words from the list of candidate term expansion words is performed by comparing between the frequencies of the words on the list of term expansion words and the frequency of the search term.

14. The method according to claim 13 wherein the frequency of a word on the list of term expansion words is the summation of all the frequencies of keywords within the associated topic that share the same stem form as the word on the list of term expansion words.

15. The method according to claim 12 further comprises expanding the search term by adding the term expansion words on the term expansion words list to the search term by using logical operators between the search term and the words on the term expansion words list.