aiwg/dist/agentic/code/frameworks/sdlc-complete/src/testing/trend-analyzer.d.ts

Deployment tool and support utility for AI context. Copies agents, skills, commands, rules, and behaviors into the paths each AI platform reads (Claude Code, Codex, Copilot, Cursor, Warp, OpenClaw, and 6 more) so one source of truth works across 10 platfo

/**
 * TrendAnalyzer - Statistical analysis for NFR trend detection
 *
 * Provides statistical methods for analyzing time-series data, detecting
 * trends, outliers, change points, and forecasting future values.
 *
 * Features:
 * - Moving average calculation (simple and exponential)
 * - Outlier detection (IQR and Z-score methods)
 * - Linear regression for trend lines
 * - Simple forecasting using trend extrapolation
 * - Change point detection using statistical tests
 *
 * @module testing/trend-analyzer
 */
/**
 * Time-series data point
 */
export interface TimeSeriesPoint {
    /** Timestamp (Unix milliseconds) */
    timestamp: number;
    /** Measured value */
    value: number;
}
/**
 * Time-series dataset
 */
export type TimeSeries = TimeSeriesPoint[];
/**
 * Linear trend line (y = mx + b)
 */
export interface TrendLine {
    /** Slope (change per unit time) */
    slope: number;
    /** Y-intercept */
    intercept: number;
    /** R-squared value (goodness of fit, 0-1) */
    rSquared: number;
}
/**
 * Forecast result with confidence
 */
export interface ForecastResult {
    /** Predicted value */
    value: number;
    /** Lower bound of prediction interval */
    lowerBound: number;
    /** Upper bound of prediction interval */
    upperBound: number;
    /** Confidence level (0-1) */
    confidence: number;
}
/**
 * Change point detection result
 */
export interface ChangePoint {
    /** Index in time series where change occurred */
    index: number;
    /** Timestamp of change point */
    timestamp: number;
    /** Mean before change */
    meanBefore: number;
    /** Mean after change */
    meanAfter: number;
    /** Statistical significance (p-value) */
    significance: number;
}
/**
 * Outlier detection method
 */
export type OutlierMethod = 'iqr' | 'zscore';
/**
 * TrendAnalyzer - Statistical analysis for NFR trends
 *
 * @example
 * ```typescript
 * const analyzer = new TrendAnalyzer();
 *
 * // Calculate moving average
 * const samples = [10, 12, 11, 13, 15, 14, 16];
 * const smoothed = analyzer.calculateMovingAverage(samples, 3);
 *
 * // Detect outliers
 * const outliers = analyzer.detectOutliers(samples, 'iqr');
 *
 * // Fit trend line
 * const timeSeries = samples.map((v, i) => ({ timestamp: i * 1000, value: v }));
 * const trend = analyzer.fitTrendLine(timeSeries);
 * console.log(`Trend: ${trend.slope > 0 ? 'increasing' : 'decreasing'}`);
 * ```
 */
export declare class TrendAnalyzer {
    /**
     * Calculate simple moving average
     *
     * @param samples - Array of values
     * @param windowSize - Size of moving window
     * @returns Array of moving averages (same length as input, padded with original values)
     * @throws {Error} If windowSize is invalid
     */
    calculateMovingAverage(samples: number[], windowSize: number): number[];
    /**
     * Calculate exponential moving average
     *
     * @param samples - Array of values
     * @param alpha - Smoothing factor (0-1, higher = more weight on recent values)
     * @returns Array of exponential moving averages
     * @throws {Error} If alpha is out of range
     */
    calculateExponentialMovingAverage(samples: number[], alpha?: number): number[];
    /**
     * Detect outliers using IQR or Z-score method
     *
     * @param samples - Array of values
     * @param method - Detection method ('iqr' or 'zscore')
     * @returns Boolean array indicating outliers (true = outlier)
     * @throws {Error} If method is invalid
     */
    detectOutliers(samples: number[], method?: OutlierMethod): boolean[];
    /**
     * Fit linear trend line to time series data
     *
     * Uses least squares regression to find best-fit line.
     *
     * @param data - Time series data points
     * @returns Trend line parameters and R-squared
     * @throws {Error} If insufficient data points
     */
    fitTrendLine(data: TimeSeries): TrendLine;
    /**
     * Forecast future value using trend extrapolation
     *
     * Uses linear regression trend line to predict future value.
     *
     * @param data - Historical time series data
     * @param horizon - Number of time units into future (milliseconds)
     * @returns Forecast with prediction interval
     * @throws {Error} If insufficient data or invalid horizon
     */
    forecastValue(data: TimeSeries, horizon: number): ForecastResult;
    /**
     * Detect change point in time series
     *
     * Uses a sliding window approach to find significant shifts in mean.
     *
     * @param data - Time series data
     * @param minSegmentSize - Minimum size of segments before/after change point
     * @returns Change point if detected, null otherwise
     */
    detectChangePoint(data: TimeSeries, minSegmentSize?: number): ChangePoint | null;
    /**
     * Calculate standard deviation of samples
     *
     * @param samples - Array of values
     * @returns Standard deviation
     */
    private calculateStdDev;
    /**
     * Calculate mean of samples
     *
     * @param samples - Array of values
     * @returns Mean value
     */
    private calculateMean;
    /**
     * Calculate percentile from sorted values
     *
     * @param sortedValues - Pre-sorted array of values
     * @param percentile - Percentile to calculate (0-100)
     * @returns Percentile value
     */
    private calculatePercentile;
    /**
     * Detect outliers using IQR method
     *
     * @private
     */
    private detectOutliersIQR;
    /**
     * Detect outliers using Z-score method
     *
     * @private
     */
    private detectOutliersZScore;
}
//# sourceMappingURL=trend-analyzer.d.ts.map