Dark Mode
IoT Intent Recognition Data
Data Science and Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
Explaining commands that are routinely given to smart home systems in order for a specified action to be executed. This dataset was originally compiled to allow experimentation with command structures and to support the eventual development of personal smart home systems, filling a gap for specific commands useful for prototyping. The collection is ideal for training machine learning algorithms focused on intent recognition, natural language processing (NLP), and categorising user interaction in Internet of Things (IoT) environments. The majority of entries involve an action, and a significant portion are phrased as questions.
Columns
- Number: An index field used for record identification.
- Category: The principal category or main theme of the command (e.g., lights, information).
- Action_needed: A binary flag indicating whether the command necessitates a physical or digital action.
- Question: A binary flag indicating if the structure of the sentence is a query or not.
- Subcategory: A more detailed grouping within the main category (e.g., specific rooms like 'kitchen').
- Action: The specific function requested to be performed (e.g., 'on', 'off', or 'none').
- Time: The requested timeframe for the action to occur, such as 'now'.
- Sentence: The natural language input or command text itself.
Distribution
The data is supplied as a CSV file (
dataset.csv) and contains 663 valid records across 8 fields. The file size is approximately 48.08 kB. All records are verified, with no mismatched or missing values detected across the columns. The content is primarily categorical and structured, with 645 unique sentence entries.Usage
This dataset can be utilised for several applications, including:
- Developing intent classification models for voice assistants or smart speakers.
- Training NLP models to segment and understand command components (action, time, object).
- Building foundational intelligent systems for home automation prototypes.
- Analysing patterns in user interactions with IoT systems.
Coverage
The scope of this collection is strictly limited to typical commands issued to smart home devices. It captures language structure related to controlling domestic appliances and requesting basic information. There is no specified geographic, time range, or demographic focus; the data represents command availability generally.
License
CC0: Public Domain
Who Can Use It
- Machine Learning Engineers: To build and validate models for multiclass classification and sequence labelling based on smart home commands.
- IoT Developers: To integrate natural language understanding functionality into new smart device platforms.
- Data Scientists: For exploratory data analysis regarding human-computer interaction patterns within domestic settings.
Dataset Name Suggestions
- Smart Home Voice Commands Log
- IoT Intent Recognition Data
- Home Automation Command Set
- User Input for Smart Systems
Attributes
Original Data Source:
QUALITYLoading...