Named entity recognition model, telephone switchboard transfer extension method and system
A named entity recognition and telephone switchboard technology, applied in the field of communication, can solve the problems of not being able to find the desired contact, affecting customer experience, reducing enterprise work efficiency, etc.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
no. 1 example
[0131] In the first embodiment, the present invention provides a named entity recognition model based on an attention-based two-way long-short-term memory unit-conditional random field (Attention-Based BiLSTM-CRF). The entity information of the model includes department names and person names, with 5 types of labels;
[0132] Wherein, the tags are: the beginning part of the person's name, the middle part of the person's name, the beginning part of the department name, the middle part of the department name and non-entity information. The tags are as follows:
[0133] The beginning of the B-Person's name
[0134] The middle part of the I-Person's name
[0135] The beginning of the B-Depart department name
[0136] The middle part of the I-Department name
[0137] ONon-entity information.
[0138] For example, "Help me transfer to Li Hong from the Information Department", after word segmentation, it is "Help me transfer to Li Hong from the Information Department", and the ou...
no. 2 example
[0161] The second embodiment is further improved on the first embodiment above, and the step of training the named entity recognition model is added, and the same part as the first embodiment above will not be repeated; as figure 2 As shown, the following steps are used to train the named entity recognition model;
[0162] S1, data preprocessing, including removing specified useless symbols, text segmentation, removing specified stop words and constructing a feature dictionary; removing specified useless symbols: redundant spaces and other meaningless symbols in the input text are useless for the model, we pre Use regular expressions to remove;
[0163] Text segmentation: jieba word segmentation, use the jieba word segmentation library to segment the text, and process the input text into a sequence of words. During the word segmentation process, for the proprietary vocabulary in some fields that may appear or the words that you do not want jieba to split, create a custom dic...
no. 3 example
[0173] In the third embodiment, the present invention provides a method for the named entity recognition model in the first or second embodiment, including the following steps:
[0174] S4, voice information to text, using the existing intelligent voice interaction platform to convert voice information into text information, such as the Alibaba Cloud intelligent voice interaction platform;
[0175] S5, extracting entity information in the text based on the named entity recognition model, including the following sub-steps;
[0176] S5.1, load the model files generated by training, including dictionaries, tags and training models;
[0177] S5.2, perform data processing on the customer's text information to generate a word index sequence; wherein, the specific steps of data processing are similar to the model training part, only need to replace the constructed dictionary with the loaded feature dictionary file;
[0178] S5.3, input the generated word index sequence into the trai...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com