Terms are the individual items that appear in the term list. Term groups are groupings of these terms. The following term groups are always available:
- Common Choices
- Content Actionability: Classifies each text into: Praise, Complaint, Suggestion, Question, Cries for Help, Safety Concerns.
- Employee Soft Skills: Bellomy has build a proprietary model that identifies mentions (direct and somewhat indirect) of employee soft skills found in the consumer comments.
- Standard Topics: The system has a standard set of topics that are modeled and classified for each text.
- Sentiment & Emotions
- Emotions – Detailed: There are 8 distinct emotions and each has 3 tiers of intensity. This group shows the 24 levels of intensity. Learn more about emotions here: Understanding Emotion Analysis
- Emotions – Simple: Shows the list of 8 main emotions. Learn more about emotions here: Understanding Emotion Analysis
- Sentiment: Each text is classified as either: Positive, Negative, Mixed, Neutral or Not Identified. Learn more here: Understanding Sentiment Analysis
- Text Extracts
- Themes: A theme is a NLP concept of an adjective-noun pair. “good show”, “terrible experience”, etc.
- Keywords & Phrases: Keyword extraction is an NLP technique that uses language rules to identify specific words and phrases in a piece of text that are sequentially related (e.g. expressed in a specific order).
- Key phrases Only: Subject of above limited to multi-word phrases.
- Any Single Word: Each word in the body of texts is shown here after some pre-processing.
- Bi-Grams: Each 2-word pair found in the body of text.
- Data Elements: Data elements are specific structured data extracted from the text. Thse include currency amounts, dates, identifiers, phone numbers, names, email addresses, URLs, hashtags and more .
- Feelings: Individual words found in the text that are commonly associated with human feelings. Note: The list of feeling words comes from academia and is not directly used for sentiment or emotion classification.
- Noun Phrases: Nouns and strings of nouns (noun-noun, adjective-noun-noun, etc.) that are found in the body of text.
- Quad Grams: Each 4-word pair found in the body of text.
- Tri-Grams: Each 3-word pair found in the body of text.
- Language Detected: Coming soon. This will indicate what language the system detected when loading the comments.
- Text Length: Lists the lengths of the text comments.
- Word Count: Lists the # of words in the text comments.
- Short Texts (<=50 Characters): If the text comment provides is <= 50 characters, they are aggregated together and listed here. Can be quite useful for analyzing short text statements.
- Supplied Variables
- VARIES – This section lists any fields that arrived with the source dataset that are reportable. In general, most close-ended questions and categorical variables will appear here.
In addition to the standard term groups, each dataset may have custom terms and term groups that are derived from the data in the dataset, or assembled by users.