Predictive and real-time adaptive music recommendation system and method
OwnerPaul Allen Blairinventor
Life of this patent
Estimating the term…
Abstract
Disclosed herein is a method for predictive and real-time adaptive music recommendation. The method comprises the steps of: acquiring data, the data comprising biometric data and user input data, the user input data indicating a context in which a playlist of music items is to be used; analyzing the data to determine a user state profile including a physiological state and a psychological state; and generating the playlist based on the user state profile. A corresponding system for predictive and real-time adaptive music recommendation is also disclosed.
Claims12
A predictive and real-time adaptive music recommendation method, comprising:
- acquiring data, the data comprising biometric data from at least one multimodal biometric sensor and user input data explicitly entered by a user via a user interface, the user input data comprising at least a mood indicator, a workout type, and a workout type intensity level, the user input data indicating a context in which a playlist of music items is to be used;
- analyzing the data to determine a user state profile, wherein analyzing comprises:
- preprocessing the biometric data to remove se ad artifacts;
- identifying key biometric metrics from the preprocessed biometric data;
- and
- jointly classifying a physiological state and a psychological state based on a combination of the key biometric metrics and the user input data, wherein the mood indicator and the workout type and workout intensity level are used together with the key biometric metrics as joint inputs to configure and perform the state classification using a machine learning algorithm;
- and
- generating the playlist based on the user state profile comprising the physiological state and the psychological state.
A predictive and real-time adaptive music recommendation system, comprising:
- a processor;
- a memory, wherein modules are stored in the memory for execution by the processor, the modules comprising:
- a data acquiring module configured to acquire data, the data comprising biometric data from at least one multimodal biometric sensor and user input data explicitly entered by a user via a user interface, the user input data comprising at least a mood indicator, a workout type, and a workout intensity level, the user input data indicating a context in which a playlist of music items is to be used;
- an analysis module configured to analyze the data to determine a user state profile, wherein the analysis module is configured to,
- preprocess the biometric data to remove noise and artifacts;
- identify key biometric metrics from the preprocessed biometric data;
- and
- jointly classify a physiological state and a psychological state based on a combination of the key biometric metrics and the user input data, wherein the mood indicator, the workout type, and the workout intensity level are used together with the key biometric metrics as joint inputs to configure and perform the state classification using a machine learning algorithm;
- and
- a recommendation module configured to generate the playlist based on the user state profile comprising the physiological state and the psychological state.
Dependencies are inferred from the claim text.