Generating a plurality of selectable responses based on a database indexed by receiver devices storing responses to similar SMS messages
A method is provided to receive an SMS message from a sender device addressed to a receiver device. Similar SMS messages addressed to the receiver device are determined. One or more of the most common responses to the similar SMS messages are determined. One or more of the most common responses are transmitted as selectable responses to the receiver device. A selected response is sent to the sender device. Another method is provided to receive an SMS message from a sender device and transmit the SMS message to a receiver device. A response from the receiver device may be received and transmit to the sender device. The SMS messages and responses collected by repeating this process may be stored in a database so that the most frequently asked questions and their corresponding most frequent answers may be determined. The most frequently asked questions and their corresponding answers may be used to build a webpage of a website.
1. A method comprising:
receiving, by at least one server, a Short Message Service (SMS) message from a sender device, wherein the SMS message is addressed to a receiver device of a plurality of receiver devices and the SMS message comprises a string of characters;
determining, by the at least one server utilizing an intent classifier, that the SMS message is requesting information and is an SMS question;
converting, by the at least one server utilizing an encoder, the SMS question into a question vector, wherein the question vector comprises a vector having at least 200 dimensions;
determining, by the at least one server utilizing a similarity engine, based on the question vector, a plurality of most similar question vectors from a plurality of receiver device question vectors stored in a database, wherein:
the plurality of receiver device question vectors are all derived from other SMS questions addressed to the receiver device,
the plurality of receiver device question vectors do not include SMS questions not addressed to the receiver device, and
the plurality of most similar question vectors are determined based on being mathematically closer to the question vector in high dimensional space as compared to other question vectors in the plurality of receiver device question vectors;
determining, by the at least one server, a plurality of most common responses connected to the plurality of most similar question vectors of a plurality of most frequently asked questions, wherein questions or closely related group of questions that have been asked the most times by a plurality of different sender devices to the same receiver device is automatically determined to be the plurality of most frequently asked questions;
transmitting, by the at least one server, the SMS message and the plurality of most common responses to the receiver device, wherein the most common responses are displayed and selectable on the receiver device;
receiving, by the at least one server, from the receiver device a selected most common response selected from the plurality of most common responses;
storing, by the at least one server, the question vector and the selected most common response in the database, wherein the question vector and the selected most common response are associated with each other and placed in a device group indexed by the receiver device; and
transmitting, by the at least one server, the selected most common response to the sender device.
2. The method of claim 1 , wherein the similarity engine uses a cosine distance or a cosine similarity to determine the plurality of most similar question vectors from the plurality of receiver device question vectors.
3. The method of claim 1 , wherein the similarity engine uses a KNN algorithm to determine which question vectors in the plurality of receiver device question vectors are the plurality of most similar question vectors.
4. The method of claim 1 , wherein the most common responses are the most common responses to the SMS question by the receiver device and not the most common responses to the SMS question by other receiver devices.
5. The method of claim 1 , wherein the SMS message and the SMS question are not stored in the database.
6. The method of claim 1 , wherein:
the database stores a plurality of question vectors, each question vector in the plurality of question vectors comprising a vector having 200 or more dimensions,
the plurality of question vectors are grouped into a plurality of device groups according to a receiver device that received the SMS message used to create the question vector, and
each question vector in the plurality of question vectors is associated with one and only one response and each response comprises a string of characters.
7. The method of claim 1 , wherein each of the plurality of receiver device question vectors is received by the same receiver device and not by any other receiver device.
8. The method of claim 1 , wherein the question vector comprises a vector having at least 10 dimensions.
9. The method of claim 1 , wherein the plurality of question vectors stored in the database are separated into a plurality of device groups and each device group in the plurality of device groups contains only question vectors created from SMS messages received from the same receiver device.
10. The method of claim 1 , wherein the plurality of most common responses are based on a history of SMS message, stored in the database, received by the receiver device.
11. A system comprising:
at least one server, storing computer code;
wherein, when executing the computer code, the at least one server is configured to:
receive a Short Message Service (SMS) message from a sender device, wherein the SMS message is addressed to a receiver device of a plurality of receiver devices and the SMS message comprises a string of characters;
convert, utilizing an encoder, the SMS message into a question vector, wherein the question vector comprises a vector having at least 200 dimensions;
determine, utilizing a similarity engine, based on the question vector, a plurality of most similar question vectors from a plurality of receiver device question vectors stored in a database, wherein:
the plurality of receiver device question vectors are all derived from other SMS questions addressed to the receiver device,
the plurality of receiver device question vectors do not include SMS questions not addressed to the receiver device, and
the plurality of most similar question vectors are determined based on being mathematically closer to the question vector in high dimensional space as compared to other question vectors in the plurality of receiver device question vectors;
determine a plurality of most common responses connected to the plurality of most similar question vectors of a plurality of most frequently asked questions, wherein questions or closely related group of questions that have been asked the most times by a plurality of different sender devices to the same receiver device is automatically determined to be the plurality of most frequently asked questions;
transmit the SMS message and the plurality of most common responses to the receiver device, wherein the most common responses are displayed and selectable on the receiver device;
receive from the receiver device a selected most common response selected from the plurality of most common responses;
store the question vector and the selected most common response in the database, wherein the question vector and the selected most common response are associated with each other and placed in a device group indexed by the receiver device; and
transmit the selected most common response to the sender device.
12. The system of claim 11 , wherein the similarity engine uses a cosine distance or a cosine similarity to determine the plurality of most similar question vectors from the plurality of receiver device question vectors.
13. The system of claim 11 , wherein the similarity engine uses a KNN algorithm to determine which question vectors in the plurality of receiver device question vectors are the plurality of most similar question vectors.
14. The system of claim 11 , wherein the most common responses are the most common responses to the SMS question by the receiver device and not the most common responses to the SMS question by other receiver devices.
15. The system of claim 11 , wherein the SMS message and the SMS question are not stored in the database.
16. The system of claim 11 , wherein:
the database stores a plurality of question vectors, each question vector in the plurality of question vectors comprising a vector having 200 or more dimensions,
the plurality of question vectors are grouped into a plurality of device groups according to a receiver device that received the SMS message used to create the question vector, and
each question vector in the plurality of question vectors is associated with one and only one response and each response comprises a string of characters.
17. The system of claim 11 , wherein each of the plurality of receiver device question vectors is received by the same receiver device and not by any other receiver device.
18. The system of claim 11 , wherein the question vector comprises a vector having at least 10 dimensions.
19. The system of claim 11 , wherein the plurality of question vectors stored in the database are separated into a plurality of device groups and each device group in the plurality of device groups contains only question vectors created from SMS messages received from the same receiver device.
20. The system of claim 11 , wherein the plurality of most common responses are based on a history of SMS message, stored in the database, received by the receiver device.