A Python library for detecting patterns and anomalies in massive datasets using the Matrix Profile

Overview

PyPI version Build Status Downloads Downloads/Week License

matrixprofile-ts

matrixprofile-ts is a Python 2 and 3 library for evaluating time series data using the Matrix Profile algorithms developed by the Keogh and Mueen research groups at UC-Riverside and the University of New Mexico. Current implementations include MASS, STMP, STAMP, STAMPI, STOMP, SCRIMP++, and FLUSS.

Read the Target blog post here.

Further academic description can be found here.

The PyPi page for matrixprofile-ts is here

Contents

Installation

Major releases of matrixprofile-ts are available on the Python Package Index:

pip install matrixprofile-ts

Details about each release can be found here.

Quick start

>>> from matrixprofile import *
>>> import numpy as np
>>> a = np.array([0.0,1.0,1.0,0.0,0.0,1.0,1.0,0.0,0.0,1.0,1.0,0.0])
>>> matrixProfile.stomp(a,4)
(array([0., 0., 0., 0., 0., 0., 0., 0., 0.]), array([4., 5., 6., 7., 0., 1., 2., 3., 0.]))

Note that SCRIMP++ is highly recommended for calculating the Matrix Profile due to its speed and anytime ability.

Examples

Jupyter notebooks containing various examples of how to use matrixprofile-ts can be found under docs/examples.

As a basic introduction, we can take a synthetic signal and use STOMP to calculate the corresponding Matrix Profile (this is the same synthetic signal as in the Golang Matrix Profile library). Code for this example can be found here

datamp

There are several items of note:

  • The Matrix Profile value jumps at each phase change. High Matrix Profile values are associated with "discords": time series behavior that hasn't been observed before.

  • Repeated patterns in the data (or "motifs") lead to low Matrix Profile values.

We can introduce an anomaly to the end of the time series and use STAMPI to detect it

datampanom

The Matrix Profile has spiked in value, highlighting the (potential) presence of a new behavior. Note that Matrix Profile anomaly detection capabilities will depend on the nature of the data, as well as the selected subquery length parameter. Like all good algorithms, it's important to try out different parameter values.

Algorithm Comparison

This section shows the matrix profile algorithms and the time it takes to compute them. It also discusses use cases on when to use one versus another. The timing comparison is based on the synthetic sample data set to show run time speed.

For a more comprehensive runtime comparison, please review the notebook docs/examples/Algorithm Comparison.ipynb.

All time comparisons were ran on a 4 core 2.8 ghz processor with 16 GB of memory. The operating system used was Ubuntu 18.04LTS 64 bit.

Algorithm Time to Complete Description
STAMP 310 ms ± 1.73 ms per loop (mean ± std. dev. of 7 runs, 1 loop each) STAMP is an anytime algorithm that lets you sample the data set to get an approximate solution. Our implementation provides you with the option to specify the sampling size in percent format.
STOMP 79.8 ms ± 473 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) STOMP computes an exact solution in a very efficient manner. When you have a historic time series that you would like to examine, STOMP is typically the quickest at giving an exact solution.
SCRIMP++ 59 ms ± 278 µs per loop (mean ± std. dev. of 7 runs, 10 loops each) SCRIMP++ merges the concepts of STAMP and STOMP together to provide an anytime algorithm that enables "interactive analysis speed". Essentially, it provides an exact or approximate solution in a very timely manner. Our implementation allows you to specify the max number of seconds you are willing to wait for a solution to obtain an approximate solution. If you are wanting the exact solution, it is able to provide that as well. The original authors of this algorithm suggest that SCRIMP++ can be used in all use cases.

Matrix Profile in Other Languages

Contact

Citations

  1. Chin-Chia Michael Yeh, Yan Zhu, Liudmila Ulanova, Nurjahan Begum, Yifei Ding, Hoang Anh Dau, Diego Furtado Silva, Abdullah Mueen, Eamonn Keogh (2016). Matrix Profile I: All Pairs Similarity Joins for Time Series: A Unifying View that Includes Motifs, Discords and Shapelets. IEEE ICDM 2016

  2. Matrix Profile II: Exploiting a Novel Algorithm and GPUs to break the one Hundred Million Barrier for Time Series Motifs and Joins. Yan Zhu, Zachary Zimmerman, Nader Shakibay Senobari, Chin-Chia Michael Yeh, Gareth Funning, Abdullah Mueen, Philip Berisk and Eamonn Keogh (2016). EEE ICDM 2016

  3. Matrix Profile V: A Generic Technique to Incorporate Domain Knowledge into Motif Discovery. Hoang Anh Dau and Eamonn Keogh. KDD'17, Halifax, Canada.

  4. Matrix Profile XI: SCRIMP++: Time Series Motif Discovery at Interactive Speed. Yan Zhu, Chin-Chia Michael Yeh, Zachary Zimmerman, Kaveh Kamgar and Eamonn Keogh, ICDM 2018.

  5. Matrix Profile VIII: Domain Agnostic Online Semantic Segmentation at Superhuman Performance Levels. Shaghayegh Gharghabi, Yifei Ding, Chin-Chia Michael Yeh, Kaveh Kamgar, Liudmila Ulanova, and Eamonn Keogh. ICDM 2017.

Owner
Target
Target's official GitHub organization
Target
Simple data balancing baselines for worst-group-accuracy benchmarks.

BalancingGroups Code to replicate the experimental results from Simple data balancing baselines achieve competitive worst-group-accuracy. Replicating

Facebook Research 29 Dec 02, 2022
Simple and flexible ML workflow engine.

This is a simple and flexible ML workflow engine. It helps to orchestrate events across a set of microservices and create executable flow to handle requests. Engine is designed to be configurable wit

Katana ML 295 Jan 06, 2023
AP1 Transcription Factor Binding Site Prediction

A machine learning project that predicted binding sites of AP1 transcription factor, using ChIP-Seq data and local DNA shape information.

1 Jan 21, 2022
pywFM is a Python wrapper for Steffen Rendle's factorization machines library libFM

pywFM pywFM is a Python wrapper for Steffen Rendle's libFM. libFM is a Factorization Machine library: Factorization machines (FM) are a generic approa

João Ferreira Loff 251 Sep 23, 2022
XGBoost + Optuna

AutoXGB XGBoost + Optuna: no brainer auto train xgboost directly from CSV files auto tune xgboost using optuna auto serve best xgboot model using fast

abhishek thakur 517 Dec 31, 2022
Uses WiFi signals :signal_strength: and machine learning to predict where you are

Uses WiFi signals and machine learning (sklearn's RandomForest) to predict where you are. Even works for small distances like 2-10 meters.

Pascal van Kooten 5k Jan 09, 2023
A Python Package to Tackle the Curse of Imbalanced Datasets in Machine Learning

imbalanced-learn imbalanced-learn is a python package offering a number of re-sampling techniques commonly used in datasets showing strong between-cla

6.2k Jan 01, 2023
A high performance and generic framework for distributed DNN training

BytePS BytePS is a high performance and general distributed training framework. It supports TensorFlow, Keras, PyTorch, and MXNet, and can run on eith

Bytedance Inc. 3.3k Dec 28, 2022
Implemented four supervised learning Machine Learning algorithms

Implemented four supervised learning Machine Learning algorithms from an algorithmic family called Classification and Regression Trees (CARTs), details see README_Report.

Teng (Elijah) Xue 0 Jan 31, 2022
The code from the Machine Learning Bookcamp book and a free course based on the book

The code from the Machine Learning Bookcamp book and a free course based on the book

Alexey Grigorev 5.5k Jan 09, 2023
Stats, linear algebra and einops for xarray

xarray-einstats Stats, linear algebra and einops for xarray ⚠️ Caution: This project is still in a very early development stage Installation To instal

ArviZ 30 Dec 28, 2022
A simple python program that draws a tree for incrementing values using the Collatz Conjecture.

Collatz Conjecture A simple python program that draws a tree for incrementing values using the Collatz Conjecture. Values which can be edited: Length

davidgasinski 1 Oct 28, 2021
Simple, fast, and parallelized symbolic regression in Python/Julia via regularized evolution and simulated annealing

Parallelized symbolic regression built on Julia, and interfaced by Python. Uses regularized evolution, simulated annealing, and gradient-free optimization.

Miles Cranmer 924 Jan 03, 2023
Feature-engine is a Python library with multiple transformers to engineer and select features for use in machine learning models.

Feature-engine is a Python library with multiple transformers to engineer and select features for use in machine learning models. Feature-engine's transformers follow scikit-learn's functionality wit

Soledad Galli 33 Dec 27, 2022
scikit-fem is a lightweight Python 3.7+ library for performing finite element assembly.

scikit-fem is a lightweight Python 3.7+ library for performing finite element assembly. Its main purpose is the transformation of bilinear forms into sparse matrices and linear forms into vectors.

Tom Gustafsson 297 Dec 13, 2022
Time Series Prediction with tf.contrib.timeseries

TensorFlow-Time-Series-Examples Additional examples for TensorFlow Time Series(TFTS). Read a Time Series with TFTS From a Numpy Array: See "test_input

Zhiyuan He 476 Nov 17, 2022
Production Grade Machine Learning Service

This project is made to help you scale from a basic Machine Learning project for research purposes to a production grade Machine Learning web service

Abdullah Zaiter 10 Apr 04, 2022
🌊 River is a Python library for online machine learning.

River is a Python library for online machine learning. It is the result of a merger between creme and scikit-multiflow. River's ambition is to be the go-to library for doing machine learning on strea

OnlineML 4k Jan 03, 2023
Generate music from midi files using BPE and markov model

Generate music from midi files using BPE and markov model

Aditya Khadilkar 37 Oct 24, 2022
fMRIprep Pipeline To Machine Learning

fMRIprep Pipeline To Machine Learning(Demo) 所有配置均在config.py文件下定义 前置环境(lilab) 各个节点均安装docker,并有fmripre的镜像 可以使用conda中的base环境(相应的第三份包之后更新) 1. fmriprep scr

Alien 3 Mar 08, 2022