Emotion early warning method, system and device and storage medium

[0055] This embodiment provides a kind of emotion early warning method, such as figure 1 As shown, the emotional early warning methods include:

[0056] Step 11. Obtain the message text corresponding to the current conversation.

[0057] Step 12: Process the message text corresponding to the current dialogue to obtain the corresponding word vector.

[0058] Specifically, it includes the data preprocessing of the message text, mainly including converting traditional Chinese to simplified, uppercase to lowercase, removing special punctuation marks, etc., and then mapping the message text into word vectors after word segmentation, so that words have semantic information. This is a commonly used means in the prior art, and will not be repeated here.

[0059] Step 13: Input the word vector into the emotion recognition model for classification to obtain the emotion recognition result corresponding to the current dialogue, and the emotion recognition model is used to recognize the emoti

[0065] This embodiment provides a kind of emotional early warning method, and this embodiment is compared with embodiment 1, and its difference is, as figure 2 As shown, the emotional early warning method also includes:

[0066] Step 10, train the deep learning model to obtain the emotion recognition model.

[0067] Such as image 3 As shown, the specific steps of step 10 in this embodiment include:

[0068] Step 101, collect sample dialog message texts.

[0069] Step 102, process the sample dialogue message text to obtain corresponding sample word vectors.

[0070] Firstly, data preprocessing is performed on the sample dialog message text, which mainly includes converting traditional Chinese to simplified Chinese, converting uppercase to lowercase, removing special punctuation marks, and word segmentation.

[0071] Each word segmentation of the sample conversation message text is mapped to a sample word vector, so that the words have semantic information.

[0072] Step 103

[0093] This embodiment provides an emotional early warning system, such as Image 6 As shown, the emotional early warning system includes an acquisition module 21 , a vector module 22 , an identification module 23 and an early warning module 24 .

[0094] The obtaining module 21 is used to obtain the message text corresponding to the current conversation.

[0095] The vector module 22 is used to process the message text corresponding to the current dialogue to obtain a corresponding word vector;

[0096] Specifically, the data preprocessing of the message text includes converting traditional Chinese to simplified Chinese, uppercase to lowercase, and removal of special punctuation marks. This is a commonly used method in the prior art and will not be described here.

[0097] After the message text is segmented, it is mapped to a word vector, so that the words have semantic information.

[0098] The recognition module 23 is used to input the word vector into the emotion recogniti