Emerging topic discovery

1. A method for discovering a new topic, the method comprising:

providing a first set of topics;

using at least one computer to:

identify a first topic within a first document;

compare the first topic to the first set of topics;

determine whether an emergence time is identified for the first topic, wherein the emergence time is based on a publishing date of at least one document, the publishing date corresponding to a date the document was first publicly available, and wherein the emergence time is identified when a quantity of appearances of the first topic in a later time period is above an appearance threshold when compared to one or more number of appearances of the topic in an earlier time period;

recognize the first topic as a new topic based on the comparison between the first topic and the first set of topics and whether an emergence time can be identified for the first topic; and

set the emergence time for the new topic to correspond to the beginning of the later time period.

2. The method of claim 1 , comprising using the at least one computer to:

identify the first set of topics from at least one document corresponding to an area of interest for topic discovery.

3. The method of claim 2 , wherein the at least one document includes at least one of a product brief, a webpage, and a technical article.

4. The method of claim 1 , comprising using the at least one computer to:

identify a plurality of topics from the first document;

determine a vector distance between the plurality of topics from the first document and the first set of topics; and

recognize the first document as related to the first set of topics based on the vector distance between the first document and the first set of documents.

5. The method of claim 4 , comprising using the at least one computer to:

generate a first vector based on which different topics occur within the plurality of topics and a frequency of each different topic in the plurality of topics; and

generate a second vector based on which different topics occur within the first set of topics and a frequency of each different topic in the first set of topics, wherein determine a vector distance includes compare the first vector to the second vector using a cosine function.

6. The method of claim 1 , wherein using the at least one computer to recognize a first topic includes using the at least one computer to identify collocated words occurring multiple times within the first document.

7. A system comprising:

a memory having software stored thereon; and

a programmable processor communicatively coupled to the memory, wherein the software, when executed by the programmable processor, causes the programmable processor to:

identify a first topic within a first document;

compare the first topic to a first set of topics;

determine whether an emergence time for the first topic is identified, wherein the emergence time is based on a publishing date of at least one document, the publishing date corresponding to a date the document was first publicly available, and wherein the emergence time is identified when a quantity of appearances of the first topic in a later time period is above an appearance threshold when compared to one or more number of appearances of the topic in an earlier time period;

recognize the first topic as a new topic based on the comparison between the first topic and the first set of topics and whether an emergence time can be identified for the first topic; and

set the emergence time for the new topic to correspond to the beginning of the later time period.

8. The system of claim 7 , wherein the software, when executed by the programmable processor, causes the programmable processor to:

identify the first set of topics from at least one document corresponding to an area of interest for topic discovery.

9. The system of claim 8 , wherein the at least one document includes at least one of a product brief, a webpage, and a technical article.

10. The system of claim 7 , wherein the software, when executed by the programmable processor, causes the programmable processor to:

identify a plurality of topics from the first document;

determine a vector distance between the plurality of topics from the first document and the first set of topics; and

recognize the first document as related to the first set of topics based on the vector distance between the first document and the first set of documents.

11. The system of claim 10 , wherein the software, when executed by the programmable processor, causes the programmable processor to:

generate a first vector based on which different topics occur within the plurality of topics and a frequency of each different topic in the plurality of topics; and

generate a second vector based on which different topics occur within the first set of topics and a frequency of each different topic in the first set of topics, wherein determine a vector distance includes compare the first vector to the second vector using a cosine function.

12. The system of claim 7 , wherein the software, when executed by the programmable processor, causes the programmable processor to:

recognize a first topic by identifying collocated words occurring multiple times within the first document.

13. A machine-readable medium including instructions which, when executed by a machine, cause the machine to perform operations comprising:

identifying a first topic within a first document;

comparing the first topic to a first set of topics;

determining whether an emergence time for the first topic is identified, wherein the emergence time is based on a publishing date of at least one document, the publishing date corresponding to a date the document was first publicly available, and wherein the emergence time is identified when a quantity of appearances of the first topic in a later time period is above an appearance threshold when compared to one or more number of appearances of the topic in an earlier time period;

recognizing the first topic as a new topic based on the comparison between the first topic and the first set of topics and whether an emergence time can be identified for the first topic; and

setting the emergence time for the new topic to correspond to the beginning of the later time period.

14. The machine readable medium of claim 13 , wherein the instructions, when executed by the machine, cause the machine to perform operations comprising:

identifying a plurality of topics from the first document;

determining a vector distance, using a cosine function, between the plurality of topics from the first document and the first set of topics; and

recognizing the first document as related to the first set of topics based on the vector distance between the first document and the first set of documents.