CO-S-01-02

Cumulative Emotion Deviation，CED

Basic Concept: The Cumulative Sentiment Deviation (CED) measures the direction and intensity of sustained sentiment deviation from its historical mean over a given period.

Core Idea:

• Z-score reflects instantaneous sentiment deviation.

• CED captures the time-accumulated effect of sentiment deviation.

1. Calculation Logic and Formula

Let the standardized sentiment momentum at time i be:

where:

• Mᵢ — Sentiment Momentum = (Positive Sentiment Ratio − Negative Sentiment Ratio)

• μᵢ, σᵢ — the mean and standard deviation within the sliding window W.

Then, at any given time t, the Cumulative Sentiment Deviation (CED) is defined as:

That is, the sentiment deviations within the most recent W observations are accumulated (summed) to quantify the persistence and direction of sentiment deviation over time.

2. Indicator Explanation

Interval	Emotional State	Market Characteristics
( CED_t ≥ +k )	Continuous Optimism Zone	Market sentiment remained positively deviated for an extended period, with a strong upward momentum in sentiment.
( -k < CED_t < +k )	Emotionally Neutral Zone	Sentiment alternated between positive and negative, with unclear direction.
( CED_t ≤ -k )	Continuous Pessimism Zone	Market sentiment remained negatively deviated for an extended period, with low confidence.

The parameter k can be set according to the historical distribution, such as ±3 or ±5, corresponding to different confidence intervals.

3. Application Guidelines

• Trend Confirmation:

  CED rising continuously → indicates bullish sentiment inertia, supporting an upward price trend.

  CED declining continuously → indicates bearish sentiment inertia, often accompanied by increased selling pressure.

• Reversal Identification: When CED reaches an extreme value and then starts to decline or reverse, it often marks a sentiment turning point in the price trend.

Official Example

• Parameter Recommendations

  Input Variable: Emotion Momentum Z-Score (standardized sentiment momentum value)

  Window Period: 30

  Typical Parameter: W = 14 (observation points)

• Example Description

• Signal Performance The abnormal threshold (k) is set to ±10 or ±15, corresponding approximately to the 2σ and 3σ ranges, respectively, for identifying extreme sentiment zones.