TrendLibrary "Trend"
calculateSlopeTrend(source, length, thresholdMultiplier)
Parameters:
source (float)
length (int)
thresholdMultiplier (float)
Purpose:
The primary goal of this function is to determine the short-term trend direction of a given data series (like closing prices). It does this by calculating the slope of the data over a specified period and then comparing that slope against a dynamic threshold based on the data's recent volatility. It classifies the trend into one of three states: Upward, Downward, or Flat.
Parameters:
`source` (Type: `series float`): This is the input data series you want to analyze. It expects a series of floating-point numbers, typically price data like `close`, `open`, `hl2` (high+low)/2, etc.
`length` (Type: `int`): This integer defines the lookback period. The function will analyze the `source` data over the last `length` bars to calculate the slope and standard deviation.
`thresholdMultiplier` (Type: `float`, Default: `0.1`): This is a sensitivity factor. It's multiplied by the standard deviation to determine how steep the slope needs to be before it's considered a true upward or downward trend. A smaller value makes it more sensitive (detects trends earlier, potentially more false signals), while a larger value makes it less sensitive (requires a stronger move to confirm a trend).
Calculation Steps:
Linear Regression: It first calculates the value of a linear regression line fitted to the `source` data over the specified `length` (`ta.linreg(source, length, 0)`). Linear regression finds the "best fit" straight line through the data points.
Slope Calculation: It then determines the slope of this linear regression line. Since `ta.linreg` gives the *value* of the line on the current bar, the slope is calculated as the difference between the current bar's linear regression value (`linRegValue`) and the previous bar's value (`linRegValue `). A positive difference means an upward slope, negative means downward.
Volatility Measurement: It calculates the standard deviation (`ta.stdev(source, length)`) of the `source` data over the same `length`. Standard deviation is a measure of how spread out the data is, essentially quantifying its recent volatility.
Adaptive Threshold: An adaptive threshold (`threshold`) is calculated by multiplying the standard deviation (`stdDev`) by the `thresholdMultiplier`. This is crucial because it means the definition of a "flat" trend adapts to the market's volatility. In volatile times, the threshold will be wider, requiring a larger slope to signal a trend. In quiet times, the threshold will be narrower.
Trend Determination: Finally, it compares the calculated `slope` to the adaptive `threshold`:
If the `slope` is greater than the positive `threshold`, the trend is considered **Upward**, and the function returns `1`.
If the `slope` is less than the negative `threshold` (`-threshold`), the trend is considered **Downward**, and the function returns `-1`.
If the `slope` falls between `-threshold` and `+threshold` (inclusive of 0), the trend is considered **Flat**, and the function returns `0`.
Return Value:
The function returns an integer representing the determined trend direction:
`1`: Upward trend
`-1`: Downward trend
`0`: Flat trend
In essence, this library function provides a way to gauge trend direction using linear regression, but with a smart filter (the adaptive threshold) to avoid classifying minor noise or low-volatility periods as significant trends.
Slope
LinearRegressionLibrary "LinearRegression"
Calculates a variety of linear regression and deviation types, with optional emphasis weighting. Additionally, multiple of slope and Pearson’s R calculations.
calcSlope(_src, _len, _condition)
Calculates the slope of a linear regression over the specified length.
Parameters:
_src (float) : (float) The source data.
_len (int) : (int) The length of the lookback period for the linear regression.
_condition (bool) : (bool) Flag to enable calculation. Set to true to calculate on every bar; otherwise, set to barstate.islast for efficiency.
Returns: (float) The slope of the linear regression.
calcReg(_src, _len, _condition)
Calculates a basic linear regression, returning y1, y2, slope, and average.
Parameters:
_src (float) : (float) The source data series.
_len (int) : (int) The length of the lookback period.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: (float ) An array of 4 values: .
calcRegStandard(_src, _len, _emphasis, _condition)
Calculates an Standard linear regression with optional emphasis.
Parameters:
_src (float) : (series float) The source data series.
_len (int) : (int) The length of the lookback period.
_emphasis (float) : (float) The emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: (float ) .
calcRegRidge(_src, _len, lambda, _emphasis, _condition)
Calculates a ridge regression with optional emphasis.
Parameters:
_src (float) : (float) The source data series.
_len (int) : (int) The length of the lookback period.
lambda (float) : (float) The ridge regularization parameter.
_emphasis (float) : (float) The emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: (float ) .
calcRegLasso(_src, _len, lambda, _emphasis, _condition)
Calculates a Lasso regression with optional emphasis.
Parameters:
_src (float) : (float) The source data series.
_len (int) : (int) The length of the lookback period.
lambda (float) : (float) The Lasso regularization parameter.
_emphasis (float) : (float) The emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: (float ) .
calcElasticNetLinReg(_src, _len, lambda1, lambda2, _emphasis, _condition)
Calculates an Elastic Net regression with optional emphasis.
Parameters:
_src (float) : (float) The source data series.
_len (int) : (int) The length of the lookback period.
lambda1 (float) : (float) L1 regularization parameter (Lasso).
lambda2 (float) : (float) L2 regularization parameter (Ridge).
_emphasis (float) : (float) Emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: (float ) .
calcRegHuber(_src, _len, delta, iterations, _emphasis, _condition)
Calculates a Huber regression using Iteratively Reweighted Least Squares (IRLS).
Parameters:
_src (float) : (float) The source data series.
_len (int) : (int) The length of the lookback period.
delta (float) : (float) Huber threshold parameter.
iterations (int) : (int) Number of IRLS iterations.
_emphasis (float) : (float) Emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: (float ) .
calcRegLAD(_src, _len, iterations, _emphasis, _condition)
Calculates a Least Absolute Deviations (LAD) regression via IRLS.
Parameters:
_src (float) : (float) The source data series.
_len (int) : (int) The length of the lookback period.
iterations (int) : (int) Number of IRLS iterations for LAD.
_emphasis (float) : (float) Emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: (float ) .
calcRegBayesian(_src, _len, priorMean, priorSpan, sigma, _emphasis, _condition)
Calculates a Bayesian linear regression with optional emphasis.
Parameters:
_src (float) : (float) The source data series.
_len (int) : (int) The length of the lookback period.
priorMean (float) : (float) The prior mean for the slope.
priorSpan (float) : (float) The prior variance (or span) for the slope.
sigma (float) : (float) The assumed standard deviation of residuals.
_emphasis (float) : (float) Emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: (float ) .
calcRFromLinReg(_src, _len, _slope, _average, _y1, _condition)
Calculates the Pearson correlation coefficient (R) based on linear regression parameters.
Parameters:
_src (float) : (float) The source data.
_len (int) : (int) The length of the lookback period.
_slope (float) : (float) The slope of the linear regression.
_average (float) : (float) The average value of the source data series.
_y1 (float) : (float) The starting point (y-intercept of the oldest bar) for the linear regression.
_condition (bool) : (bool) Flag to enable calculation. Set to true to calculate on every bar; otherwise, set to barstate.islast for efficiency.
Returns: (float) The Pearson correlation coefficient (R) adjusted for the direction of the slope.
calcRFromSource(_src, _len, _condition)
Calculates the correlation coefficient (R) using a specified length and source data.
Parameters:
_src (float) : (float) The source data.
_len (int) : (int) The length of the lookback period.
_condition (bool) : (bool) Flag to enable calculation. Set to true to calculate on every bar; otherwise, set to barstate.islast for efficiency.
Returns: (float) The correlation coefficient (R).
calcSlopeLengthZero(_src, _len, _minLen, _step, _condition)
Identifies the length at which the slope is flattest (closest to zero).
Parameters:
_src (float) : (float) The source data.
_len (int) : (int) The maximum lookback length to consider (minimum of 2).
_minLen (int) : (int) The minimum length to start from (cannot exceed the max length).
_step (int) : (int) The increment step for lengths.
_condition (bool) : (bool) Flag to enable calculation. Set to true to calculate on every bar; otherwise, set to barstate.islast.
Returns: (int) The length at which the slope is flattest.
calcSlopeLengthHighest(_src, _len, _minLen, _step, _condition)
Identifies the length at which the slope is highest.
Parameters:
_src (float) : (float) The source data.
_len (int) : (int) The maximum lookback length (minimum of 2).
_minLen (int) : (int) The minimum length to start from.
_step (int) : (int) The step for incrementing lengths.
_condition (bool) : (bool) Flag to enable calculation. Set to true to calculate on every bar; otherwise, set to barstate.islast.
Returns: (int) The length at which the slope is highest.
calcSlopeLengthLowest(_src, _len, _minLen, _step, _condition)
Identifies the length at which the slope is lowest.
Parameters:
_src (float) : (float) The source data.
_len (int) : (int) The maximum lookback length (minimum of 2).
_minLen (int) : (int) The minimum length to start from.
_step (int) : (int) The step for incrementing lengths.
_condition (bool) : (bool) Flag to enable calculation. Set to true to calculate on every bar; otherwise, set to barstate.islast.
Returns: (int) The length at which the slope is lowest.
calcSlopeLengthAbsolute(_src, _len, _minLen, _step, _condition)
Identifies the length at which the absolute slope value is highest.
Parameters:
_src (float) : (float) The source data.
_len (int) : (int) The maximum lookback length (minimum of 2).
_minLen (int) : (int) The minimum length to start from.
_step (int) : (int) The step for incrementing lengths.
_condition (bool) : (bool) Flag to enable calculation. Set to true to calculate on every bar; otherwise, set to barstate.islast.
Returns: (int) The length at which the absolute slope value is highest.
calcRLengthZero(_src, _len, _minLen, _step, _condition)
Identifies the length with the lowest absolute R value.
Parameters:
_src (float) : (float) The source data.
_len (int) : (int) The maximum lookback length (minimum of 2).
_minLen (int) : (int) The minimum length to start from.
_step (int) : (int) The step for incrementing lengths.
_condition (bool) : (bool) Flag to enable calculation. Set to true to calculate on every bar; otherwise, set to barstate.islast.
Returns: (int) The length with the lowest absolute R value.
calcRLengthHighest(_src, _len, _minLen, _step, _condition)
Identifies the length with the highest R value.
Parameters:
_src (float) : (float) The source data.
_len (int) : (int) The maximum lookback length (minimum of 2).
_minLen (int) : (int) The minimum length to start from.
_step (int) : (int) The step for incrementing lengths.
_condition (bool) : (bool) Flag to enable calculation. Set to true to calculate on every bar; otherwise, set to barstate.islast.
Returns: (int) The length with the highest R value.
calcRLengthLowest(_src, _len, _minLen, _step, _condition)
Identifies the length with the lowest R value.
Parameters:
_src (float) : (float) The source data.
_len (int) : (int) The maximum lookback length (minimum of 2).
_minLen (int) : (int) The minimum length to start from.
_step (int) : (int) The step for incrementing lengths.
_condition (bool) : (bool) Flag to enable calculation. Set to true to calculate on every bar; otherwise, set to barstate.islast.
Returns: (int) The length with the lowest R value.
calcRLengthAbsolute(_src, _len, _minLen, _step, _condition)
Identifies the length with the highest absolute R value.
Parameters:
_src (float) : (float) The source data.
_len (int) : (int) The maximum lookback length (minimum of 2).
_minLen (int) : (int) The minimum length to start from.
_step (int) : (int) The step for incrementing lengths.
_condition (bool) : (bool) Flag to enable calculation. Set to true to calculate on every bar; otherwise, set to barstate.islast.
Returns: (int) The length with the highest absolute R value.
calcDevReverse(_src, _len, _slope, _y1, _inputDev, _emphasis, _condition)
Calculates the regressive linear deviation in reverse order, with optional emphasis on recent data.
Parameters:
_src (float) : (float) The source data.
_len (int) : (int) The length of the lookback period.
_slope (float) : (float) The slope of the linear regression.
_y1 (float) : (float) The y-intercept (oldest bar) of the linear regression.
_inputDev (float) : (float) The input deviation multiplier.
_emphasis (float) : (float) Emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: A 2-element tuple: .
calcDevForward(_src, _len, _slope, _y1, _inputDev, _emphasis, _condition)
Calculates the progressive linear deviation in forward order (oldest to most recent bar), with optional emphasis.
Parameters:
_src (float) : (float) The source data array, where _src is oldest and _src is most recent.
_len (int) : (int) The length of the lookback period.
_slope (float) : (float) The slope of the linear regression.
_y1 (float) : (float) The y-intercept of the linear regression (value at the most recent bar, adjusted by slope).
_inputDev (float) : (float) The input deviation multiplier.
_emphasis (float) : (float) Emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: A 2-element tuple: .
calcDevBalanced(_src, _len, _slope, _y1, _inputDev, _emphasis, _condition)
Calculates the balanced linear deviation with optional emphasis on recent or older data.
Parameters:
_src (float) : (float) Source data array, where _src is the most recent and _src is the oldest.
_len (int) : (int) The length of the lookback period.
_slope (float) : (float) The slope of the linear regression.
_y1 (float) : (float) The y-intercept of the linear regression (value at the oldest bar).
_inputDev (float) : (float) The input deviation multiplier.
_emphasis (float) : (float) Emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: A 2-element tuple: .
calcDevMean(_src, _len, _slope, _y1, _inputDev, _emphasis, _condition)
Calculates the mean absolute deviation from a forward-applied linear trend (oldest to most recent), with optional emphasis.
Parameters:
_src (float) : (float) The source data array, where _src is the most recent and _src is the oldest.
_len (int) : (int) The length of the lookback period.
_slope (float) : (float) The slope of the linear regression.
_y1 (float) : (float) The y-intercept (oldest bar) of the linear regression.
_inputDev (float) : (float) The input deviation multiplier.
_emphasis (float) : (float) Emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: A 2-element tuple: .
calcDevMedian(_src, _len, _slope, _y1, _inputDev, _emphasis, _condition)
Calculates the median absolute deviation with optional emphasis on recent data.
Parameters:
_src (float) : (float) The source data array (index 0 = oldest, index _len - 1 = most recent).
_len (int) : (int) The length of the lookback period.
_slope (float) : (float) The slope of the linear regression.
_y1 (float) : (float) The y-intercept (oldest bar) of the linear regression.
_inputDev (float) : (float) The deviation multiplier.
_emphasis (float) : (float) Emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns:
calcDevPercent(_y1, _inputDev, _condition)
Calculates the percent deviation from a given value and a specified percentage.
Parameters:
_y1 (float) : (float) The base value from which to calculate deviation.
_inputDev (float) : (float) The deviation percentage.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: A 2-element tuple: .
calcDevFitted(_len, _slope, _y1, _emphasis, _condition)
Calculates the weighted fitted deviation based on high and low series data, showing max deviation, with optional emphasis.
Parameters:
_len (int) : (int) The length of the lookback period.
_slope (float) : (float) The slope of the linear regression.
_y1 (float) : (float) The Y-intercept (oldest bar) of the linear regression.
_emphasis (float) : (float) Emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: A 2-element tuple: .
calcDevATR(_src, _len, _slope, _y1, _inputDev, _emphasis, _condition)
Calculates an ATR-style deviation with optional emphasis on recent data.
Parameters:
_src (float) : (float) The source data (typically close).
_len (int) : (int) The length of the lookback period.
_slope (float) : (float) The slope of the linear regression.
_y1 (float) : (float) The Y-intercept (oldest bar) of the linear regression.
_inputDev (float) : (float) The input deviation multiplier.
_emphasis (float) : (float) Emphasis factor: 0 for equal weight; >0 emphasizes recent bars; <0 emphasizes older bars.
_condition (bool) : (bool) Flag to enable calculation (true = calculate).
Returns: A 2-element tuple: .
calcPricePositionPercent(_top, _bot, _src)
Calculates the percent position of a price within a linear regression channel. Top=100%, Bottom=0%.
Parameters:
_top (float) : (float) The top (positive) deviation, corresponding to 100%.
_bot (float) : (float) The bottom (negative) deviation, corresponding to 0%.
_src (float) : (float) The source price.
Returns: (float) The percent position within the channel.
plotLinReg(_len, _y1, _y2, _slope, _devTop, _devBot, _scaleTypeLog, _lineWidth, _extendLines, _channelStyle, _colorFill, _colUpLine, _colDnLine, _colUpFill, _colDnFill)
Plots the linear regression line and its deviations, with configurable styles and fill.
Parameters:
_len (int) : (int) The lookback period for the linear regression.
_y1 (float) : (float) The starting y-value of the regression line.
_y2 (float) : (float) The ending y-value of the regression line.
_slope (float) : (float) The slope of the regression line (used to determine line color).
_devTop (float) : (float) The top deviation to add to the line.
_devBot (float) : (float) The bottom deviation to subtract from the line.
_scaleTypeLog (bool) : (bool) Use a log scale if true; otherwise, linear scale.
_lineWidth (int) : (int) The width of the plotted lines.
_extendLines (string) : (string) How lines should extend (none, left, right, both).
_channelStyle (string) : (string) The style of the channel lines (solid, dashed, dotted).
_colorFill (bool) : (bool) Whether to fill the space between the top and bottom deviation lines.
_colUpLine (color) : (color) Line color when slope is positive.
_colDnLine (color) : (color) Line color when slope is negative.
_colUpFill (color) : (color) Fill color when slope is positive.
_colDnFill (color) : (color) Fill color when slope is negative.
Slope_TKLibrary "Slope_TK"
This library calculate the slope of a serie between two points
The serie can be ta.ema(close,200) for example
The size is the number of bars between the two points for the slope calculation, for example it can be 10
slope_of_ema200 = slope(t a.eam(close, 200) , 10 )
slope( float serie, int size )
LinearRegressionLibraryLibrary "LinearRegressionLibrary" contains functions for fitting a regression line to the time series by means of different models, as well as functions for estimating the accuracy of the fit.
Linear regression algorithms:
RepeatedMedian(y, n, lastBar) applies repeated median regression (robust linear regression algorithm) to the input time series within the selected interval.
Parameters:
y :: float series, source time series (e.g. close)
n :: integer, the length of the selected time interval
lastBar :: integer, index of the last bar of the selected time interval (defines the position of the interval)
Output:
mSlope :: float, slope of the regression line
mInter :: float, intercept of the regression line
TheilSen(y, n, lastBar) applies the Theil-Sen estimator (robust linear regression algorithm) to the input time series within the selected interval.
Parameters:
y :: float series, source time series
n :: integer, the length of the selected time interval
lastBar :: integer, index of the last bar of the selected time interval (defines the position of the interval)
Output:
tsSlope :: float, slope of the regression line
tsInter :: float, intercept of the regression line
OrdinaryLeastSquares(y, n, lastBar) applies the ordinary least squares regression (non-robust) to the input time series within the selected interval.
Parameters:
y :: float series, source time series
n :: integer, the length of the selected time interval
lastBar :: integer, index of the last bar of the selected time interval (defines the position of the interval)
Output:
olsSlope :: float, slope of the regression line
olsInter :: float, intercept of the regression line
Model performance metrics:
metricRMSE(y, n, lastBar, slope, intercept) returns the Root-Mean-Square Error (RMSE) of the regression. The better the model, the lower the RMSE.
Parameters:
y :: float series, source time series (e.g. close)
n :: integer, the length of the selected time interval
lastBar :: integer, index of the last bar of the selected time interval (defines the position of the interval)
slope :: float, slope of the evaluated linear regression line
intercept :: float, intercept of the evaluated linear regression line
Output:
rmse :: float, RMSE value
metricMAE(y, n, lastBar, slope, intercept) returns the Mean Absolute Error (MAE) of the regression. MAE is is similar to RMSE but is less sensitive to outliers. The better the model, the lower the MAE.
Parameters:
y :: float series, source time series
n :: integer, the length of the selected time interval
lastBar :: integer, index of the last bar of the selected time interval (defines the position of the interval)
slope :: float, slope of the evaluated linear regression line
intercept :: float, intercept of the evaluated linear regression line
Output:
mae :: float, MAE value
metricR2(y, n, lastBar, slope, intercept) returns the coefficient of determination (R squared) of the regression. The better the linear regression fits the data (compared to the sample mean), the closer the value of the R squared is to 1.
Parameters:
y :: float series, source time series
n :: integer, the length of the selected time interval
lastBar :: integer, index of the last bar of the selected time interval (defines the position of the interval)
slope :: float, slope of the evaluated linear regression line
intercept :: float, intercept of the evaluated linear regression line
Output:
Rsq :: float, R-sqared score
Usage example:
//@version=5
indicator('ExampleLinReg', overlay=true)
// import the library
import tbiktag/LinearRegressionLibrary/1 as linreg
// define the studied interval: last 100 bars
int Npoints = 100
int lastBar = bar_index
int firstBar = bar_index - Npoints
// apply repeated median regression to the closing price time series within the specified interval
{square bracket}slope, intercept{square bracket} = linreg.RepeatedMedian(close, Npoints, lastBar)
// calculate the root-mean-square error of the obtained linear fit
rmse = linreg.metricRMSE(close, Npoints, lastBar, slope, intercept)
// plot the line and print the RMSE value
float y1 = intercept
float y2 = intercept + slope * (Npoints - 1)
if barstate.islast
{indent} line.new(firstBar,y1, lastBar,y2)
{indent} label.new(lastBar,y2,text='RMSE = '+str.format("{0,number,#.#}", rmse))