System and method for identifying sentiment in text strings

1. A computer-implemented method comprising:

receiving an input;

evaluating the input with a first model to determine an initial sentiment and confidence thereof;

if the confidence is greater than a threshold, using the initial sentiment;

if the confidence is lower than the threshold, accessing a list of secondary sentiments and evaluating the input for each of the secondary sentiments;

if the confidence is lower than the threshold, generating a relevantly similar confidence (RSC) score associated with each of the secondary sentiments; and

if the confidence is lower than the threshold, determining whether to use the initial sentiment or one sentiment of the secondary sentiments based on the generated RSC scores.

2. The computer-implemented method of claim 1 , wherein the method further comprises:

selecting a sentiment based on the determining; and

displaying the selected sentiment associated with the input.

3. The computer-implemented method of claim 1 , wherein the list of secondary sentiments is a client specific posting list comprising a plurality of correction inputs, wherein each RSC score is based on a determination of relevance and similarity of the input to each correction input of the plurality of correction inputs.

4. The computer-implemented method of claim 3 , wherein each correction input of the plurality of correction inputs has been previously provided by a user.

5. The computer-implemented method of claim 1 , wherein when the RSC score is null or less than another threshold, the initial sentiment is selected as a resolved sentiment.

6. The computer-implemented method of claim 1 , wherein when only one RSC score is equal to or greater than another threshold, the secondary sentiment corresponding to that RSC score equal to or greater than the another threshold is selected as a resolved sentiment.

7. The computer-implemented method of claim 1 , wherein when multiple RSC scores are equal to or greater than another threshold, the method further comprises using secondary sentiments corresponding to the multiple RSC scores that are equal to or greater than the another threshold as input factors to select a tertiary sentiment as a resolved sentiment.

8. The computer-implemented method of claim 7 , further comprising:

receiving a user input as the tertiary sentiment.

9. The computer-implemented method of claim 1 , wherein the list of the secondary sentiments is associated with one client, and wherein a relevance similar analysis model used to generate the RSC score only accesses the list of secondary sentiments associated with the one client when generating the RSC scores for the input for the one client.

10. The computer-implemented method of claim 1 , wherein the initial sentiment and the RSC scores are generated in real-time.

11. The computer-implemented method of claim 1 , wherein the list of secondary sentiments is populated by client administered entries.

12. The computer-implemented method of claim 1 , wherein the input is a limited text input.

13. The computer-implemented method of claim 1 , wherein a relevance similar analysis model used to generate the RSC scores is trained with a dataset used by a first model, and wherein the relevance similar analysis model uses the dataset and the list of secondary sentiments to evaluate the input to generate the RSC scores.

14. A computer-implemented method comprising:

receiving an input at an application server from at least one remotely-connected computing device over a wide-area network;

transmitting the input from the application server to a remotely-connected analysis server for analysis; and

receiving analysis result from the analysis server, wherein the analysis result is based on running an initial sentiment model and determining an initial confidence score associated with the initial sentiment model, and wherein the analysis result is a selection of the initial sentiment based on an initial sentiment model and output of the initial sentiment if the initial confidence score is above a predetermined threshold, and wherein the analysis result is based on another sentiment model and another confidence score associated therewith if the initial confidence score is below the predetermined threshold and wherein the analysis result is a selection and output by a resolution module between the initial sentiment based on the initial sentiment model and an another sentiment based on the another sentiment model.

15. The computer-implemented method of claim 14 , wherein the analysis result is received by the application server over a network.

16. The computer-implemented method of claim 14 further comprising sending data associated with the analysis result from the application server to the at least one remotely-connected computing device over the wide-area network.

17. The computer-implemented method of claim 16 , wherein the data is rendered on a display of the at least one remotely-connected computing device, wherein the rendition includes a plurality of user responses received from the at least one remotely-connected computing device, and wherein each user response of the plurality of user responses includes a text string that was a basis of the analysis result by the analysis server.

18. The computer-implemented method of claim 14 , wherein the initial sentiment model is a universal sentiment analysis model.