Time series analysis is a powerful tool that can be used to extract valuable information from data and make predictions about future events. It can be used to identify trends, seasonal patterns, and other relationships between variables. Time series analysis can also be used to predict future events, such as changes in sales, demand, or prices.
If you're working with time series data in Python, there are many different libraries you can use. In this article, we'll review the most popular libraries in Python in no particular order.
1、Sktime
Sktime is a Python toolkit for working with time series data. It provides a set of tools for working with time series data, including tools for manipulating, visualizing, and analyzing the data. Sktime is designed to be easy to use and extensible, so new time series algorithms can be easily implemented.
Sktime provides extensions to the scikit-learn API. It includes all the necessary methods and tools for efficiently solving problems involving time series regression, forecasting, and classification. The library contains specialized machine learning algorithms and transformation methods for time series. These are features not available in any other tool library.
sktime also provides interfaces to related libraries, such as scikit-learn, statsmodels, tsfresh, PyOD, and fbprophet, etc.
example:
from sktime.datasets import load_airline
from sktime.forecasting.model_selection import temporal_train_test_split
# from sktime.utils.plotting.forecasting import plot_ys
y = load_airline()
y_train, y_test = temporal_train_test_split(y)
plt.title('Airline Data with Train and Test')
y_train.plot(label = 'train')
y_test.plot(label = 'test')
plt.legend()2、pmdarima
pmdarima is a Python library for statistical analysis of time series data. It is based on ARIMA models and provides a variety of tools for analyzing, forecasting, and visualizing time series data. Pmdarima also provides a variety of tools for working with seasonal data, including seasonal testing and seasonal decomposition tools.
One of the forecasting models often used in time series analysis is ARIMA (Autoregressive Integrated Moving Average). ARIMA is a forecasting algorithm whereby we can predict future values based on information in the past values of a time series without any additional information.
pmdarima is a wrapper for ARIMA models and comes with an automatic function to automatically find the best hyperparameters (p,d,q) for an arima model. The library includes and features the following:
-
Equivalent to the R function auto.arima
-
A collection of statistical tests for stationarity and seasonality
-
Time series utilities such as differencing and inverse differencing
-
Many endogenous and exogenous transformers and featureizers, including Box-Cox and Fourier transforms
-
Seasonal Time Series Decomposition
-
Cross Validation Utility
-
Rich built-in time series datasets for prototyping and examples
To learn more about this library, check out this link:https://github.com/alkaline-ml/pmdarima
3、tsfresh
tsfresh is a Python package that automates the process of feature extraction from time series. It is based on the idea that information in a time series can be decomposed into a set of meaningful features called features. tsfresh takes care of the tedious work of manually extracting these features and provides tools for automatic feature selection and classification. This package is designed to be used with pandas DataFrames and provides a wide range of functions for working with time series data, including:
-
Automatic feature extraction from time series
-
automatic feature selection
-
time series decomposition
-
Dimensionality reduction
-
Outlier Detection
-
Supports multiple time series formats
-
Support for missing values
-
Support multiple languages
To learn more about this library, check out this link:https://github.com/blue-yonder/tsfresh
4、Prophet
Facebook's Prophet is a forecasting tool that anyone can consume with data in CSV format. Prophet is open source software released by Facebook's core data science team. It is based on an additive model in which non-linear trends are fitted with annual, weekly and daily seasonality and holiday effects. It works best for time series with strong seasonal effects and multiple seasons of historical data. Prophet is robust to missing data and changes in trend, and generally handles outliers well.
According to the official documentation, fbprophet works well with time series data with significant seasonal effects and several seasons of previous data. Additionally, fbprophet claims that it is resistant to missing data and manages outliers efficiently.
To learn more about this library, check out this link:https://github.com/facebook/prophet
5、State forecast
Statsforecast provides a collection of widely used univariate time series forecasting models, it includes a large number of benchmark models:
-
Fastest and most accurate in .AutoARIMAPythonR
-
Fastest and most accurate in .ETSPythonR
-
Replace FB-Prophet with two lines of code, improve speed and accuracy.
-
Compatible with sklearn interface.
-
Include and for ARIMA.exogenous variables prediction intervals
-
Compile into high-performance machine code through .numba
6、PyCaret
PyCaret is an open-source low-code machine learning library in Python that automates machine learning workflows. It is an end-to-end machine learning and model management tool that exponentially speeds up experimentation cycles and increases your productivity.
Compared to other open source machine learning libraries, PyCaret is an alternative low-code library that can be used to replace hundreds of lines of code with only a few lines of code. This makes experiments exponentially faster and more efficient. PyCaret is essentially a Python wrapper around several machine learning libraries and frameworks, such as scikit-learn, XGBoost, LightGBM, CatBoost, spaCy, Optuna, Hyperopt, Ray, and many more.
Although PyCaret is not a dedicated time series forecasting library, it has a new module dedicated to time series forecasting. It's still in pre-release mode, but you can try it out by installing the tagged pycaret.
The PyCaret time series module is consistent with the existing API and fully functional. Statistical testing, model training and selection (over 30 algorithms), model analysis, automatic hyperparameter tuning, experiment recording, cloud deployment, etc. All this with just a few lines of code.
pip install --pre pycaret
To learn more about this library, check out this link:https://github.com/pycaret/pycaret
in conclusion
There are many time series forecasting libraries available in Python (more than I can cover here). Each has its own strengths and weaknesses, so it's important to choose the right one for your needs.