A Navigation Voice POI Understanding and Correction Method and System Based on User Profiles

Abstract

This invention relates to a navigation voice POI understanding and correction method and system based on user profiles. The method includes: constructing a POI history database and a POI negative feedback database based on the user's historical navigation data; collecting the user's voice POIs; transmitting the voice POIs and recognizing them according to the POI history database and the POI negative feedback database; and using the corresponding standard voice POI as the result. The navigation voice POI understanding and correction method of this invention constructs a POI history database and a POI negative feedback database based on the user's historical navigation data. After collecting the user's voice POIs, it searches and judges them in the POI history database and the POI negative feedback database, recognizes the voice POIs, and finds the corresponding standard voice POIs as the output result for navigation. This method achieves personalized understanding of voice POIs through retrieval and judgment in the POI history database and the POI negative feedback database, improving the accuracy and recall rate of voice POI recognition, thereby improving user experience and user satisfaction.

Claims

1. A navigation voice POI understanding and correction method based on user profiles, characterized in that, include: Based on users' historical navigation data, a POI history library and a POI negative feedback library are constructed. Specifically, based on users' historical navigation data, a mapping with the voice POI list is generated to construct the POI history library. The POI information of the incomplete navigation corresponding to each user's voice POI is extracted to construct a set list, and the POI voice text and the standard POI list are mapped to construct the POI negative feedback library. Collect user's voice POI; Transmit voice POIs and recognize them based on the POI history database and the POI negative feedback database; The corresponding standard speech POI after recognition is taken as the result; The process of collecting user voice POIs also includes: calculating the user's preference score S for each POI based on the user's historical navigation data, and constructing a user profile system. The preference score S is calculated as follows: Where u represents the user, i represents the POI, S(u,i) represents the user's preference score for the POI, N represents the number of times the user has navigated to the same POI in the past year, and W c Based on online experience, the weighting constant for different POI navigation sources is represented by t. j F(t) represents the number of days since the user completed navigation to the POI. j ) represents the time decay factor, and λ represents the exponential time constant; The system transmits voice navigation information and identifies POIs based on a historical POI database and a negative feedback POI database. This includes: acquiring the associated historical POI database, performing text retrieval and recall, filtering POI information below a similarity threshold, and generating a candidate POI list with high relevance to the voice POIs; determining whether there are POIs in the negative feedback POI database in the candidate POI list, filtering them out if they are, generating a valid candidate POI list, and sorting them according to similarity scores, while also sorting POIs with similarity scores by preference scores.

2. The navigation voice POI understanding and correction method according to claim 1, characterized in that, The construction of the POI history database and POI negative feedback database includes: generating different user IDs according to different users, and constructing a POI history database and POI negative feedback database corresponding to each user ID.

3. The navigation voice POI understanding and correction method according to claim 1, characterized in that, The process of collecting users' voice POIs also includes: building a POI abbreviation and alias library based on users' historical navigation data.

4. The navigation voice POI understanding and correction method according to claim 3, characterized in that, Transmitting voice navigation information and identifying it based on the POI history database and the POI negative feedback database includes: retrieving the mapped POI from the POI abbreviation and alias database; if the mapped POI is found, determining whether it is in the negative feedback list; if not, using the POI as the result; if it is, filtering based on the POI negative feedback database and re-identifying based on the POI history database.

5. The navigation voice POI understanding and correction method according to claim 1, characterized in that, The system transmits voice navigation information and identifies POIs based on a historical POI database and a negative feedback POI database. This includes: acquiring the associated historical POI database, performing text retrieval and retrieval, filtering POI information below a similarity threshold, generating a candidate POI list with high relevance to the voice POIs, determining whether there are POIs in the negative feedback POI database in the candidate POI list, filtering them out if they are, generating a valid candidate POI list, and sorting them according to similarity scores.

6. The navigation voice POI understanding and correction method according to claim 5, characterized in that, The standard speech POI is the POI information with the highest similarity score.

7. The navigation voice POI understanding and correction method according to any one of claims 1-6, characterized in that, The process of collecting user voice POIs includes: collecting voice POIs, converting voice POIs into text information, and extracting voice POIs from the text information.

8. A navigation voice POI understanding and correction system based on user profiles, characterized in that, The system is used to implement the navigation voice POI understanding and correction method according to any one of claims 1-6; the system includes an in-vehicle voice device and a cloud server that interacts with the in-vehicle voice device, wherein the cloud server is provided with a POI data module, a negative feedback mechanism module and a POI understanding module for voice POI recognition.

Images