Skip to content

nb module

Numba-compiled functions for label generation.

Note

Set wait to 1 to exclude the current value from calculation of future values.

Warning

Do not attempt to use these functions for building predictor variables as they may introduce the look-ahead bias to your model - only use for building target variables.

bin_trend_labels_1d_nb function

bin_trend_labels_1d_nb(
    pivots
)

Values classified into 0 (downtrend) and 1 (uptrend).

bin_trend_labels_nb function

bin_trend_labels_nb(
    pivots
)

2-dim version of bin_trend_labels_1d_nb().

binc_trend_labels_1d_nb function

binc_trend_labels_1d_nb(
    high,
    low,
    pivots
)

Median values normalized between 0 (downtrend) and 1 (uptrend).

binc_trend_labels_nb function

binc_trend_labels_nb(
    high,
    low,
    pivots
)

2-dim version of binc_trend_labels_1d_nb().

bincs_trend_labels_1d_nb function

bincs_trend_labels_1d_nb(
    high,
    low,
    pivots,
    up_th,
    down_th
)

Median values normalized between 0 (downtrend) and 1 (uptrend) but capped once the threshold defined at the beginning of the trend is exceeded.

bincs_trend_labels_nb function

bincs_trend_labels_nb(
    high,
    low,
    pivots,
    up_th,
    down_th
)

2-dim version of bincs_trend_labels_1d_nb().

breakout_labels_1d_nb function

breakout_labels_1d_nb(
    high,
    low,
    window=14,
    up_th=inf,
    down_th=inf,
    wait=1
)

For each value, return 1 if any value in the next period is greater than the positive threshold (in %), -1 if less than the negative threshold, and 0 otherwise.

First hit wins. Continue search if both thresholds were hit at the same time.

breakout_labels_nb function

breakout_labels_nb(
    high,
    low,
    window=14,
    up_th=inf,
    down_th=inf,
    wait=1
)

2-dim version of breakout_labels_1d_nb().

fixed_labels_1d_nb function

fixed_labels_1d_nb(
    close,
    n=1
)

Percentage change of the current value relative to a future value.

fixed_labels_nb function

fixed_labels_nb(
    close,
    n=1
)

2-dim version of fixed_labels_1d_nb().

future_max_1d_nb function

future_max_1d_nb(
    close,
    window=14,
    wait=1,
    minp=None
)

Rolling maximum over future values.

future_max_nb function

future_max_nb(
    close,
    window=14,
    wait=1,
    minp=None
)

2-dim version of future_max_1d_nb().

future_mean_1d_nb function

future_mean_1d_nb(
    close,
    window=14,
    wtype=0,
    wait=1,
    minp=None,
    adjust=False
)

Rolling average over future values.

future_mean_nb function

future_mean_nb(
    close,
    window=14,
    wtype=0,
    wait=1,
    minp=None,
    adjust=False
)

2-dim version of future_mean_1d_nb().

future_min_1d_nb function

future_min_1d_nb(
    close,
    window=14,
    wait=1,
    minp=None
)

Rolling minimum over future values.

future_min_nb function

future_min_nb(
    close,
    window=14,
    wait=1,
    minp=None
)

2-dim version of future_min_1d_nb().

future_std_1d_nb function

future_std_1d_nb(
    close,
    window=14,
    wtype=0,
    wait=1,
    minp=None,
    adjust=False,
    ddof=0
)

Rolling standard deviation over future values.

future_std_nb function

future_std_nb(
    close,
    window=14,
    wtype=0,
    wait=1,
    minp=None,
    adjust=False,
    ddof=0
)

2-dim version of future_std_1d_nb().

iter_symmetric_down_th_nb function

iter_symmetric_down_th_nb(
    up_th
)

Negative upper threshold that is symmetric to a positive one at one iteration.

iter_symmetric_up_th_nb function

iter_symmetric_up_th_nb(
    down_th
)

Positive upper threshold that is symmetric to a negative one at one iteration.

For example, 50% down requires 100% to go up to the initial level.

mean_labels_1d_nb function

mean_labels_1d_nb(
    close,
    window=14,
    wtype=0,
    wait=1,
    minp=None,
    adjust=False
)

Percentage change of the current value relative to the average of a future period.

mean_labels_nb function

mean_labels_nb(
    close,
    window=14,
    wtype=0,
    wait=1,
    minp=None,
    adjust=False
)

2-dim version of mean_labels_1d_nb().

pct_trend_labels_1d_nb function

pct_trend_labels_1d_nb(
    high,
    low,
    pivots,
    normalize=False
)

Percentage change of median values relative to the next pivot.

pct_trend_labels_nb function

pct_trend_labels_nb(
    high,
    low,
    pivots,
    normalize=False
)

2-dim version of pct_trend_labels_1d_nb().

pivots_1d_nb function

pivots_1d_nb(
    high,
    low,
    up_th,
    down_th
)

Pivots denoted by 1 (peak), 0 (no pivot) or -1 (valley).

Two adjacent peak and valley points should exceed the given threshold parameters.

If any threshold is given element-wise, it will be applied per new/updated pivot.

pivots_nb function

pivots_nb(
    high,
    low,
    up_th,
    down_th
)

2-dim version of pivots_1d_nb().

trend_labels_1d_nb function

trend_labels_1d_nb(
    high,
    low,
    up_th,
    down_th,
    mode=0
)

Trend labels based on TrendLabelMode.

trend_labels_nb function

trend_labels_nb(
    high,
    low,
    up_th,
    down_th,
    mode=0
)

2-dim version of trend_labels_1d_nb().