Mood Prediction Model Using Duration-Preprocessed Sequence Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for predicting psychological states from user behavior data fail to accurately capture time-oriented fluctuations in mood and emotion, and do not provide insights into the duration of predicted mood states, limiting their effectiveness in understanding mental health influences.
Innovation Solution
A learning device that preprocesses psychological state data to include duration information, and learns a psychological state sequence prediction model using both behavior sequence data and preprocessed mood data, enabling the prediction of future mood sequences along with their duration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional methods use daily mean values for mood prediction, then the prediction process is simplified, but time-oriented fluctuations in mood cannot be captured
Solution Approach 1:
The patent segments mood data by time of day into multiple periods (e.g., morning, afternoon, evening), allowing the system to capture intra-daily mood fluctuations. Each period is processed separately to generate period-specific predictions, thereby preserving temporal variations that would be lost in daily mean aggregation.
Solution Approach 2:
The patent introduces a temporal dimension by dividing the day into multiple time periods. Instead of a single daily aggregate value, the system creates a multi-dimensional structure where mood is analyzed across different time slots, enabling capture of diurnal patterns and fluctuations.
2Device complexity
If future mood data sequence is predicted without duration information, then the prediction model is simpler, but the time interval of mood fluctuations cannot be determined
Solution Approach 1:
The patent performs preliminary processing to calculate mood duration by identifying consecutive data points with the same mood label. This pre-computed duration information is then integrated into the prediction model as an additional feature, enabling the model to learn temporal patterns without requiring complex post-processing.
Solution Approach 2:
The patent introduces duration as an intermediary variable that bridges the gap between raw mood sequences and prediction outcomes. By calculating and incorporating duration information, the system enables the prediction model to understand temporal characteristics of mood states without directly modeling complex time intervals.
3Ease of manufacture
If mood data is aggregated to mean values, then data processing is easier, but individual mood variations are lost
Solution Approach 1:
The patent segments mood data into multiple time-period-specific datasets rather than aggregating into a single daily mean. This segmentation preserves individual mood variations within each period while still enabling systematic processing, as each segment can be handled independently with standardized procedures.
Data Source
AI summary
A learning device includes: a psychological state data preprocessing unit that calculates a duration of a psychological state from psychological state sequence data, and generates preprocessed psychological state sequence data including the psychological state and the duration; and a learning unit that learns a psychological state sequence prediction model, using input sequence data including behavior sequence data and the preprocessed psychological state sequence data, and correct sequence data that is preprocessed psychological state sequence data at a time later than the input sequence data.