Adaptive theme extraction from communication data in a communication center environment
A system for theming is disclosed. A transformer receives a plurality of sampled thought objects and a prompt from a theming computer. The prompt is designed to provide context data, format requirements, and instructions for theme assignment. The plurality of thought objects comprises text present in data from communication environments and the prompt requests the transformer to assign themes to each thought object. An object-theming transformer with a list of themes determines the probability score for mapping a thought object to a list of themes. When an associated probability score is above a threshold, the thought object is assigned to a known theme. In an embodiment, un-themed thought objects (not mapped to known themes) are provided to a topic identification transformer to generate theme names. The themes assigned and/or generated by the theming transformer and topic identification transformer are displayed on a graphical user interface.
1. A system for theming a plurality of thought objects, the system comprising:
a theming computer comprising one or more processors, a memory, a plurality of programming instructions stored in the memory that are executed by the one or more processors, the instructions cause the processor to:
communicate with one or more transformers, wherein the one or more transformers are configured to:
receive a plurality of thought objects and a prompt, wherein the plurality of thought objects comprises text inputs present in data from communication environments, and wherein the prompt comprises a request to determine themes for each thought object;
for each thought object, determine, using a theming transformer of the one or more transformers, whether a probability score is above a pre-defined threshold, wherein the probability score is indicative of the thought object mapping to a known theme;
responsive to the probability score of the thought object being above the pre-defined threshold, associate a known theme to the thought object, wherein the known theme is among previously themed thought objects;
responsive to the probability score of the thought object being below the pre-defined threshold, generate, using a topic identification transformer of the transformers, a theme names of un-themed thought objects based on topics identified;
wherein generating each theme name comprises:
providing un-themed thought objects to the topic identification transformer, wherein the topic identification transformer is configured to:
identify topics in the un-themed thought objects;
determine a number of themes for the un-themed thought objects;
label identified topics into a theme; and
assign the theme to the thought object; and
assign and display the themes associated with the plurality of thought objects on a graphical user interface of a user device.
2. The system of claim 1 , wherein the plurality of programming instructions further cause the processor to:
provide a list of themes to the theming transformer;
generate a probability score between 0 and 1 indicating likelihood of each thought object mapping to each known theme; and
return results in a structured format associating thought objects to the known theme with corresponding probability score.
3. The system of claim 1 , wherein the received plurality of thought objects is generated by the theming computer using random sampling or stratified sampling.
4. The system of claim 1 , wherein the prompt comprises context data, format requirements, and instructions for theme assignment.
5. The system of claim 1 , wherein the prompt comprises a prompt assignment configured to cause the theming transformer to determine probability scores for mapping thought objects to known themes.
6. The system of claim 1 , wherein the prompt comprises either:
a default prompt configured for initial topic identification; or
a stable prompt configured to maintain consistent themes across multiple analyses.
7. A computer-implemented method for assigning themes to a plurality of thought objects, the method comprising the steps of:
receiving, at one or more transformers, a plurality of thought objects and a prompt, wherein the plurality of thought objects comprises text inputs present in data from communication environments, and wherein the prompt comprises a request to the one or more transformers to determine themes to the plurality of thought objects;
for each thought object, determining, using a theming transformer of the one or more transformers, whether a probability score is above a pre-defined threshold, wherein the probability score is indicative of the thought object mapping to a known theme;
responsive to the probability score of the thought object being above the pre-defined threshold, associating a known theme with the thought object, wherein the known theme is among previously themed thought objects;
responsive to the probability score of the thought object being below the pre-defined threshold, generating, using a topic identification transformer of the one or more transformers, theme names of un-themed thought objects based on topics identified;
wherein generating each theme name comprises:
providing un-themed thought objects to the topic identification transformer, wherein the topic identification transformer is configured to:
identifying topics in the un-themed thought objects;
determining a number of themes for the un-themed thought objects;
labeling identified topics into a theme; and
assigning the theme to the thought object; and
assigning and displaying the themes associated with the plurality of thought objects on a graphical user interface of a user device.
8. The computer-implemented method of claim 7 , wherein receiving the prompt at the theming transformer further comprises the steps of:
receiving a list of themes at the one or more transformers;
receiving instructions for generating a probability score between 0 and 1 indicating a likelihood of the thought object mapping to each known theme; and
receiving instructions to return results in a structured format associating thought objects to the known theme with corresponding probability scores.
9. The computer-implemented method of claim 7 , wherein the received plurality of thought objects is generated by the theming computer using simple random sampling or stratified sampling.
10. The computer-implemented method of claim 7 , wherein the prompt comprises context data, format requirements, and instructions for theme assignment.
11. The computer-implemented method of claim 7 , wherein the prompt comprises a prompt assignment configured to cause the theming transformer to determine probability scores for mapping thought objects to known themes.
12. The computer-implemented method of claim 7 , wherein the prompt comprises either:
a default prompt configured for initial topic identification; or
a stable prompt configured to maintain consistent themes across multiple analyses.