Upload sample training sentences
Last updated
Last updated
Step 1: Select the Sample Sentences tab (1).
Step 2: Click Upload sample sentence (2), then select click to download example file (3) to download the input sample file to upload to the system.
Step 3: Enter the sample sentence, intent, and corresponding entity data file.
Note
1) Intent name: ask_working_hour (Separate words with an underscore).
2) Entity: Saturday |sys.date (separate between entity and entity type with a dash). If the sentence has multiple entities, separate them with semicolons.
3) Do not use special characters to name intents or entities (For example: (!@#$%^&*.)
4) Can use English or Vietnamese without accents to name intentions and entities.
Step 4: After entering data, select Upload sample sentence (5) then click Select file (6) to upload the sample sentence to be trained.
Step 5: Select Processing (7) to upload the training data file to the system.
In case the system already has data available, the trainer selects Replace to delete all old data and replace it with uploaded data.
Step 6: After confirming, the system processes the training file. The results include details of the error sentence (if any) and the error data line (8). The trainer makes necessary data corrections.
Step 1: Select the Sample Sentences tab (1) then click the Create New button (2).
Step 2: At the interface, select Intent (3) in the pre-created list. (See also Intent section )
Step 3: Enter a sample sentence corresponding to the selected intention in the Content section.
Step 4: Highlight the word you want to enter the entity type in the sample sentence. Select the corresponding entity type from the list available in the system or enter a new entity type (4). (See also Entity section )
Step 5: Select Save (5) to complete entering the sample sentence, intent, and entity.
Coaches can add new intents on the Intents tab as follows:
Step 1: Select the Context tab (1) then click the Create New button.
Step 2: Enter the intent name in the Intent tab and add the necessary descriptions to the intent name (2).
Step 3: Click Save (3) new intention.
At this interface, the trainer can add sample sentences corresponding to the following intent:
Step 1: At the Samples tab, click the Create New button (1). In the displayed dialog section, enter the content of the sample sentence corresponding to the intention (2) and entity contained in the sample sentence (3).
Step 2: Select the Create new button (4) to complete the step of creating a sample sentence.
Trainers can add entity-type data at the Entity tab as follows:
Step 1: At the Entity Type tab (1), click Create New (2). At the display interface, enter the entity type content, and description (Content section) and select the corresponding Entity Type (3).
In the entity type section, there are similar options
For example: Consider the sample sentence "Is the Dien Bien Phu branch open tomorrow at 7 am? " This sample sentence intends to ask for working hours (ask_working_hour: date) with two entity information appearing together in the sentence that needs to be extracted: Saturday (entity type: date) and Dien Bien Phu branch (type entity: branch_name).
Step 2: Click Save (4) of the newly created entity type.
Step 3: After creating a new entity, the trainer selects the Training button in the toolbar to train new knowledge for the Bot.
System entities are pre-built entities on the EM&AI platform.
System entities can help identify basic information such as date (sys.date), city (sys.city), quantity (sys.amount), time period (sys.duration), email (sys.email)...
When the trainer assigns the information contained in the customer's sentence to the corresponding system entity type, the Bot can identify and extract that information without needing to be trained through many sample sentences.
Step 1: Select Entity List -> Select Create New.
Step 2: At the display interface, enter the main keyword content into the Entity List section, and enter the corresponding acronym, synonym,... of the main keyword into the Synonyms section. Select the entity type that needs to be recognized for the keyword from the Available Entity Types list.
Step 3: Click Save (5) to save the new entity list.
Step 1: Select Entity List -> select Upload -> Select Click to download sample keywords.
Step 2: The trainer enters keywords into the sample data file in the following format.
In there:
Keyword: The main keyword that needs to be identified by the system.
Synonym: Synonyms/Abbreviations/Borrowed words/... have the same meaning as the main keyword.
Entity Type: Entity type of that keyword.
Step 3: Click Select File(4) then select the entered keyword file. Click Processing (5) for the system to start uploading keywords to the system. The results show the number of error/no error sentences after uploading.
Stopwords là những từ có trong từ điển tiếng Việt nhưng không ảnh hưởng đến nghĩa. Huấn luyện Stopwords giúp hệ thống nhận diện được những từ này để xử lý loại bỏ nếu xuất hiện trong câu. Chọn tab Stopwords (1), tại giao diện hiển thị sau đó nhập danh sách từ stopwords (2) cần xử lý vào ô Nội dung.
Stopwords are words that do not affect the meaning of the utterance. Stopwords training helps the system identify these words for processing and removal if they appear in a sentence. Select the Stopwords (1) tab, at the display interface then enter the list of stopwords (2) words that need to be processed in the Content section.
Step 1: Select the Teencodes tab.
Step 2: Enter the meaning you need to recognize for teencodes in the "Enter value" section (2) and teencodes in the "Codes" section (3). The trainer can add multiple corresponding teencodes values, then press Enter.
Step 3: Click Add (4) to save Teencodes.
Normal
Composite:
Common entity type.
Extract multiple information that appears together in a sentence.
