System and methods for using conversational similarity for dimension reduction in deep analytics
A system for using conversational similarity for dimension reduction in deep analytics, comprising a self-learning interaction optimizer that receives string-based data from a contact center and analyzes it to produce a plurality of information similarity vectors, provides the vectors to a neural network and receives output vectors from the neural network, and produces context data from the output vectors and associates the context data with the original string-based data.
1. A system that uses for using conversational similarity to reduce dimensional complexity within contact center analysis operations for dimension reduction in deep analytics, comprising:
a self-learning interaction optimizer comprising at least a plurality of programming instructions stored in a memory and operating on a processor of a computing, wherein the programmable instructions, when operating on the processor, cause the processor to:
observe and analyze string-based data from a contact center system via a network, wherein a reinforcement learning server communicates with the contact center in order to access the string-based data in real-time as streaming data during runtime and an event analyzer interprets received events to take next actions that are to be executed;
produce a plurality of information similarity vectors based at least in part on the analysis, wherein the similarity vectors describe the occurrences of each word within specified proximity based on a similarity strength of the word;
provide at least a portion of the information similarity vectors as input variables to a dilated convolutional artificial neural network to generate a lower-dimensional dense vector;
receive a plurality of output vectors as lower-dimensional dense vectors from the dilated convolutional artificial neural network;
produce context data based at least in part on the output vectors; and
associate at least a portion of the context data with at least a portion of the string-based data corresponding to the input vectors;
wherein the self-learning interaction optimizer transmits at least a portion of the string-based data and the corresponding associated context data as output to a plurality of contact center resources, the contact center resources comprising a routing server; and
wherein the routing server routes an inbound interaction to a destination contact center resource based at least in part on the received context data associated with string-based data.
2. The system of claim 1 , wherein the analysis comprises a comparison of word occurrences within at least two strings within the string-based data.
3. The system of claim 2 , wherein the analysis further comprises a comparison of word occurrence proximity within the at least two strings.
4. The system of claim 1 , wherein the analysis comprises a comparison of occurrences of at least a character substring within at least two strings within the string-based data.
5. The system of claim 4 , wherein the character substring comprises a character trigram comprising exactly three characters.
6. A method for using conversational similarity for dimension reduction in deep analytics, comprising the steps of:
receiving, at a self-learning interaction optimizer comprising at least a plurality of programming instructions stored in a memory and operating on a processor of a computing device, string-based data from a contact center system via a network;
observing and analyzing at least a portion of the string-based data from a contact center system via a network, wherein a reinforcement learning server communicates with the contact center in order to access the string-based data in real-time as streaming data during runtime and an event analyzer interprets received events to take next actions that are to be executed;
producing a plurality of information similarity vectors based at least in part on the analysis, wherein the similarity vectors describe the occurrences of each word within specified proximity based on a similarity strength of the word;
providing at least a portion of the information similarity vectors as input variables to a dilated convolutional artificial neural network to generate a lower-dimensional dense vector;
receiving a plurality of output vectors as lower-dimensional dense vectors from the dilated convolutional artificial neural network;
producing context data based at least in part on the output vectors; and
associating at least a portion of the context data with at least a portion of the string-based data corresponding to the input vectors;
wherein the self-learning interaction optimizer transmits at least a portion of the string-based data and the corresponding associated context data as output to a plurality of contact center resources, the contact center resources comprising a routing server: and
wherein the routing server routes an inbound interaction to a destination contact center resource based at least in part on the received context data associated with string-based data.
7. The method of claim 6 , wherein the analysis comprises a comparison of word occurrences within at least two strings within the string-based data.
8. The method of claim 7 , wherein the analysis further comprises a comparison of word occurrence proximity within the at least two strings.
9. The method of claim 6 , wherein the analysis comprises a comparison of occurrences of at least a character substring within at least two strings within the string-based data.
10. The method of claim 9 , wherein the character substring comprises a character trigram comprising exactly three characters.