Standard entity text determination method and device based on BiLSTM model and storage medium
A determination method and entity technology, applied to devices and storage media, in the field of standard entity text determination methods based on the BiLSTM model, can solve the problems of low applicability, low accuracy and low efficiency, achieve high consistency rate and improve efficiency , the effect of improving the accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0055] In the candidate entity recall stage, the candidate entity set is constructed based on the literal similarity of strings, text statistical features, and Elasticsearch search engine retrieval. Text matching at this stage is only equivalent to coarse screening. Non-standard text data has a variety of expressions. Different words in Chinese texts may have the same meaning, and some diagnostic original words with similar expressions have differences in word order. The text matching methods used in the coarse screening process have low accuracy and cannot meet our needs, so they can also be matched with semantic similarity. In the candidate entity disambiguation stage, using text semantic matching information can improve the quality of entity normalization.
[0056] At present, there are two main frameworks for semantic similarity matching based on deep learning. One is the Siamese twin network, which uses a parameter-sharing symmetric network to model input entity pairs, an...
Embodiment 2
[0139] Figure 9 It is a block diagram of a standard entity text determination device based on the BiLSTM model provided by the embodiment of the present invention. In this embodiment, the device is applied to figure 1 A standard entity text determination method based on the BiLSTM model is shown. The device includes at least the following modules:
[0140] A selection module 51, configured to select a corresponding candidate entity set for the received text entity to be matched;
[0141] The grouping module 52 is used for forming a text entity pair with the text entity to be matched for each candidate entity in the candidate entity set;
[0142] The feature vector module 53 is used for each text entity pair, using the preset neural matching neural network to calculate the first similarity feature vector of the text entity pair, and using the text statistical method and the fully connected network to calculate the second similarity of the text entity pair degree feature vec...
Embodiment 3
[0148] The embodiment of the present invention provides a standard entity text determination device based on the BiLSTM model, which is used for determining a standard entity text based on the BiLSTM model, such as Figure 10 As shown, the electronic device includes a processor 1001 and a memory 1002, wherein the processor 1001 and the memory 1002 can be connected through a bus or other methods, Figure 10 Take connection via bus as an example.
[0149] The processor 1001 can be a central processing unit (Central Processing Unit, CPU) or other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), graphics processing units (Graphics Processing Unit, GPU), embedded neural network processing Neural-network Processing Unit (NPU) or other dedicated deep learning coprocessor, Application Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (Field-Programmable Gate Array, FPGA) or other programmable logic Chips such as devices, discrete ...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More