System and method for electronic chat production

1. A computer program product comprising a non-transitory, computer-readable medium embodying thereon a set of computer instructions, the set of computer instructions comprising instructions for:

accessing an electronic chat, the electronic chat embodying a set of electronic chat messages;

adaptively splitting the set of electronic chat messages into a set of conversations, each conversation in the set of conversations comprising a subset of electronic chat messages from the set of electronic chat messages, wherein adaptively splitting the set of electronic chat messages into a set of conversations further comprises:

determining a set of time gaps between adjacent messages from the set of electronic chat messages;

learning, using the set of time gaps, a Gaussian mixture model representing a mixture of Gaussian distributions;

determining a highest mean value distribution from the mixture of Gaussian distributions;

identifying a plurality of split points based on the determined highest mean value distribution from the mixture of Gaussian distributions; and storing each conversation from the set of conversations as a separate document.

2. The computer program product of claim 1 , wherein each electronic chat message embodied in the electronic chat has associated metadata and wherein adaptively splitting the set of electronic chat messages into the set of conversations comprises clustering the set of electronic chat messages into clusters based on the associated metadata of the electronic chat messages from the set of electronic chat messages.

3. The computer program product of claim 1 , wherein each electronic chat message embodied in the electronic chat has a timestamp and wherein adaptively splitting the set of electronic chat messages into the set of conversations comprises clustering the set of electronic chat messages into clusters based on the timestamps of the electronic chat messages from the set of electronic chat messages.

4. The computer program product of claim 1 , wherein each electronic chat message embodied in the electronic chat has a timestamp, and wherein the set of computer instructions comprises instructions for:

determining a set of models that model the set of time gaps, wherein determining the set of models comprises:

determining a single Gaussian distribution of the set of time gaps;

selecting a best model from the set of models;

based on selecting the single Gaussian distribution as the best model, not splitting the electronic chat; and

based on selecting the Gaussian mixture model as the best model, performing the adaptive splitting of the set of electronic chat messages into the set of conversations based on the Gaussian mixture model.

5. The computer program product of claim 4 , wherein selecting the best model from the set of models comprises determining, for each model in the set of models, a Bayesian information criterion and selecting the best model from the set of models based on the Bayesian information criteria for the set of models.

6. The computer program product of claim 4 , wherein performing the adaptive splitting of the set of electronic chat messages into the set of conversations based on the Gaussian mixture model comprises:

selecting a time gap from the set of time gaps:

determining a probability of the selected time gap for each Gaussian distribution in the mixture of Gaussian distributions to produce a set of probabilities for the selected time gap; and

based on a determination that a highest probability from the set of probabilities for the selected time gap is for the highest mean value distribution, splitting the electronic chat into new conversation at the selected time gap.

7. The computer program product of claim 6 , wherein the set of computer instructions comprises instructions for not splitting the electronic chat at the selected time gap based on a determination that the highest probability from the set of probabilities for the selected time gap is not for the highest mean value distribution.

8. An computer-implemented method comprising:

receiving, by a computer processor, an electronic chat, the electronic chat embodying a set of electronic chat messages;

adaptively splitting, by the computer processor, the set of electronic chat messages into a set of conversations, each conversation in the set of conversations comprising a subset of electronic chat messages from the set of electronic chat messages, wherein adaptively splitting the set of electronic chat messages into a set of conversations further comprises:

determining a set of time gaps between adjacent messages from the set of electronic chat messages;

learning, using the set of time gaps, a Gaussian mixture model representing a mixture of Gaussian distributions;

determining a highest mean value distribution from the mixture of Gaussian distributions;

identifying a plurality of split points based on the determined highest mean value distribution from the mixture of Gaussian distributions; and

storing, by the computer processor, each conversation from the set of conversations as a separate document.

9. The computer-implemented method of claim 8 , wherein each electronic chat message embodied in the electronic chat has associated metadata and wherein adaptively splitting the set of electronic chat messages into the set of conversations comprises clustering the set of electronic chat messages into clusters based on the associated metadata of the electronic chat messages from the set of electronic chat messages.

10. The computer-implemented method of claim 8 , wherein each electronic chat message embodied in the electronic chat has a timestamp and wherein adaptively splitting the set of electronic chat messages into the set of conversations comprises clustering the set of electronic chat messages into clusters based on the timestamps of the electronic chat messages from the set of electronic chat messages.

11. The computer-implemented method of claim 8 , further comprising:

determining, by the computer processor, a set of models that model the set of time gaps, wherein determining the set of models comprises:

determining a single Gaussian distribution of the set of time gaps;

selecting, by the computer processor, a best model from the set of models;

based on selecting the Gaussian mixture model as the best model, performing the adaptive splitting of the set of electronic chat messages into the set of conversations based on the Gaussian mixture model.

12. The computer-implemented method of claim 8 , wherein selecting the best model from the set of models comprises determining a Bayesian information criterion for each model in the set of models and selecting the best model from the set of models based on the Bayesian information criteria for the set of models.

13. The computer-implemented method of claim 11 , wherein performing the adaptive splitting of the set of electronic chat messages into the set of conversations based on the Gaussian mixture model comprises:

selecting a time gap from the set of time gaps:

determining a probability of the selected time gap for each Gaussian distribution in the mixture of Gaussian distributions to produce a set of probabilities for the selected time gap; and

based on a determination that a highest probability from the set of probabilities for the selected time gap is for the highest mean value distribution, splitting the electronic chat into new conversation at the selected time gap.

14. An e-discovery system comprising:

a computer processor;

a non-transitory, computer-readable medium embodying thereon a set of computer instructions executable by the computer processor, the set of computer instructions comprising instructions for:

sending a chat query to a remote electronic chat service;

receiving an electronic chat responsive to the chat query, the electronic chat embodying a set of electronic chat messages;

adaptively splitting the set of electronic chat messages into a set of conversations, each conversation in the set of conversations comprising a subset of electronic chat messages from the set of electronic chat messages, wherein adaptively splitting the set of electronic chat messages into a set of conversations further comprises:

determining a set of time gaps between adjacent messages from the set of electronic chat messages;

learning, using the set of time gaps, a Gaussian mixture model representing a mixture of Gaussian distributions;

determining a highest mean value distribution from the mixture of Gaussian distributions;

identifying a plurality of split points based on the determined highest mean value distribution from the mixture of Gaussian distributions; and

storing each conversation from the set of conversations as a separate document.

15. The e-discovery system of claim 14 , wherein each electronic chat message embodied in the electronic chat has associated metadata and wherein adaptively splitting the set of electronic chat messages into the set of conversations comprises clustering the set of electronic chat messages into clusters based on the associated metadata of the electronic chat messages from the set of electronic chat messages.

16. The e-discovery system of claim 14 , wherein each electronic chat message embodied in the electronic chat has a timestamp and wherein adaptively splitting the set of electronic chat messages into the set of conversations comprises clustering the set of electronic chat messages into clusters based on the timestamps of the electronic chat messages from the set of electronic chat messages.

17. The e-discovery system of claim 14 , wherein each electronic chat message embodied in the electronic chat has a timestamp, and wherein the set of computer instructions comprises instructions for:

determining a set of models that model the set of time gaps, wherein determining the set of models comprises:

determining a single Gaussian distribution of the set of time gaps;

selecting a best model from the set of models;

based on selecting the single Gaussian distribution as the best model, not splitting the electronic chat; and

based on selecting the Gaussian mixture model as the best model, performing the adaptive splitting of the set of electronic chat messages into the set of conversations based on the Gaussian mixture model.

18. The e-discovery system of claim 17 , wherein selecting the best model from the set of models comprises determining, for each model in the set of models, a Bayesian information criterion and selecting the best model from the set of models based on the Bayesian information criteria for the set of models.

19. The e-discovery system of claim 17 , wherein performing the adaptive splitting of the set of electronic chat messages into the set of conversations based on the Gaussian mixture model comprises:

selecting a time gap from the set of time gaps:

determining a probability of the selected time gap for each Gaussian distribution in the mixture of Gaussian distributions to produce a set of probabilities for the selected time gap; and

based on a determination that a highest probability from the set of probabilities for the selected time gap is for the highest mean value distribution, splitting the electronic chat into new conversation at the selected time gap.

20. The e-discovery system of claim 19 , wherein the set of computer instructions comprises instructions for not splitting the electronic chat at the selected time gap based on a determination that the highest probability from the set of probabilities for the selected time gap is not for the highest mean value distribution.

21. The e-discovery system of claim 14 , further comprising a search engine, wherein each conversation from the set of conversations is separately indexable for searching by the search engine.

22. The e-discovery system of claim 14 , further comprising a machine learning classifier, wherein each conversation from the set of conversations is separately classifiable by the machine learning classifier.