Skip to content

Library for automated signal segmentation, trend classification and analysis.

License

Notifications You must be signed in to change notification settings

izikeros/trend_classifier

Repository files navigation

Trend classifier

pre-commit.ci status Black formatter flake8 pytest Maintainability codecov

Library for automated signal segmentation, trend classification and analysis.

Installation

  1. The package is pip-installable. To install it, run:

    pip3 install trend-classifier

Usage

Pandas DataFrame Input

usage:

import yfinance as yf
from trend_classifier import Segmenter

# download data from yahoo finance
df = yf.download("AAPL", start="2018-09-15", end="2022-09-05", interval="1d", progress=False)

x_in = list(range(0, len(df.index.tolist()), 1))
y_in = df["Adj Close"].tolist()

seg = Segmenter(x_in, y_in, n=20)
seg.calculate_segments()

For graphical output use Segmenter.plot_segments():

seg.plot_segments()

Segmentation example

After calling method Segmenter.calculate_segments() segments are identified and information is stored in Segmenter.segments as list of Segment objects. Each Segment object. Each Segment object has attributes such as 'start', 'stop' - range of indices for the extracted segment, slope and many more attributes that might be helpful for further analysis.

Exemplary info on one segment:

from devtools import debug
debug(seg.segments[3])

and you should see something like this:

    seg.segments[3]: Segment(
        start=154,
        stop=177,
        slope=-0.37934038908585044,
        offset=109.54630934894907,
        slopes=[
            -0.45173184100846725,
            -0.22564684358754555,
            0.15555037018051593,
            0.34801127785130714,
        ],
        offsets=[
            121.65628807526804,
            83.56079272220015,
            17.32660986821478,
            -17.86417581658647,
        ],
        slopes_std=0.31334199799377654,
        offsets_std=54.60900279722876,
        std=0.933497081795997,
        span=82.0,
        reason_for_new_segment='offset',
    )

export results to tabular format (pandas DataFrame):

seg.segments.to_dataframe()

(NOTE: for clarity reasons, not all columns are shown in the screenshot above)

Alternative approach

  • Smooth out the price data using the Savitzky-Golay filter,
  • label the highs and lows.
  • higher highs and higher lows indicates an uptrend.

The requirement here is than you need OHLC data for the assets you would like to analyse.

License

MIT © Krystian Safjan.