{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Supervised learning\n", "\n", "Supervised ML is the task of learning a function that maps an input to an output based on example input-output pairs. \n", "Formally, we are given with a set $\\mathcal{D}$ consists of (data,labels) pairs: \n", "$$\\mathcal{D} = \\{ (x_i , y_i) \\}_{i=1}^{m}$$ \n", "\n", "where $x_i \\in \\mathcal{X}$ are the datapoints and $y_i \\in \\mathcal{Y}$ are the labels. For simplicity, we assume here that the \"labels space\" $\\mathcal{Y}$ is a finite set $y_i$ that are discrete, univariate variables, i.e., classification settings.\n", "\n", "The goal in supervised learning is to **fit** a function $f : \\mathcal{X} \\to \\mathcal{Y}$ such that $f(x_i) =y_i$ for all $i=1,\\dots,m$\\ .
\n", "Traditionaly, the data points $x_i$ are elements of some *vector space*\\, meaning that, each point can be expressed using a $p$-tuple (vector) of numbers\n", "$$x_i = (x_{i1}, x_{i2}, \\dots, x_{ip})$$\n", "\n", "hence we have $f : \\mathbb{R}^p \\to \\mathcal{Y}$. \n", "However, when working with multi-way data, and time-series data in particular, the situation is different. \n", "\n", "## The case of matrix valued inputs\n", "\n", "To adhere with the empirical results and demonstrations shown in Mor et al., we restrict the discussion to the case where data points are gathered across multiple timepoints. That is, each sample $x_i$ is in fact, an $n$ by $p$ matrix, where each of the $p$ columns represents a feature and the rows correspond to different timepoints in which the samples were gathered.\n", "\n", "\n", "The matrix input setting is not uncommon, for example, consider the task of gray-scale image classification, where each image is a set of $n \\times p$ pixels and each pixel holds a value from 0 to 1 determining its brightness.
\n", "In the case of longitudinal sampling data, we can think of the rows of each sample $x_i$ as discrete samples of a \"curve\" in a $p$ dimensional space.
\n", "In formal words, let $\\gamma_i : \\mathbb{R} \\to \\mathbb{R}^p$ a smooth function, and $\\{t_j\\}_{j=1}^n \\subset \\mathbb{R}$ be a set of timepoints, such that $\\gamma_i (t_j) \\in \\mathbb{R}^p$ is the $j^{th}$ row of the matrix $x_i$\\ . " ] }, { "cell_type": "raw", "metadata": { "raw_mimetype": "text/restructuredtext" }, "source": [ ".. figure:: ./data/single_curve_demo.png\n", " :alt: curve\n", " :class: with-shadow\n", " :width: 50%\n", " :align: center\n", "\n", " An illustration of the :math:`p` dimensional curve :math:`\\gamma_i` (here :math:`p=3`\\ ), were the blue line is the curve itself, points illustrate discrete sampling process, each point correspond to a timepoint :math:`t_j` for :math:`j=1,\\dots ,n`\\ ." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In our context, we are given with collection of such labeled curves, all of which are sampled in corresponding timepoints, and our goal is to fit a function that takes a sampled curve $x =[\\gamma (t_1), \\gamma (t_2) , \\dots , \\gamma (t_n)]$ and outputs the label associated with the curve." ] }, { "cell_type": "raw", "metadata": { "raw_mimetype": "text/restructuredtext" }, "source": [ ".. figure:: ./data/multiple_curves_demo.png\n", " :alt: curves\n", " :class: with-shadow\n", " :width: 50%\n", " :align: center\n", "\n", " Illustration of the supervised curve classification task, curves are colored according to their class label. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "\n", "\n", "## ML workflows\n", "\n", "The function $f: \\mathbb{R}^{n \\times p} \\to \\mathcal{Y}$ usually consists of a composition of \"simpler\" functions \n", "$$f(x) = f_1 \\circ f_2 \\circ ... \\circ f_d (x)$$ \n", "\n", "i.e., a *pipeline* of functions, where each function $f_j$ \"captures\" a certain property of the data that is key for making an accurate prediction, and such that the mapping $f_j$ is vector valued, starting some $j \\leq d$\\ . Meaning that $f_k (x) \\in \\mathbb{R}^{\\ell_k}$ for all $k\\geq j$\\ . \n", "Examples for common functions of use in supevised ML are \\\n" ] }, { "cell_type": "raw", "metadata": { "raw_mimetype": "text/restructuredtext" }, "source": [ "* \\ In neural networks, :math:`f_j` are of the form :math:`f_j (x) = \\alpha_j ({\\bf W}_j x + b_j)` where \n", "\n", " * \\ :math:`{\\bf W}_j` is a matrix representing a :math:`\\mathbb{R}^{j_1} \\to \\mathbb{R}^{j_2}` linear mapping \n", " * \\ :math:`b_j \\in \\mathbb{R}^{j_2}` is a bias term \n", " * \\ :math:`\\alpha_j` is a non-linear activation function such as, :math:`\\tanh, \\operatorname{SoftMax} , \\operatorname{sigmoid}` etc... \n", " \n", "* \\ Normalization steps, such as standard-scaling \n", "* \\ Dimensionality reduction steps, like PCA, CCA, UMAM, etc..." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "An essential requirement we impose on the functions $f_j$ is that after the training process is done, and the parameters are learned, each $f_j$ must be applicable to samples outside of the training set $\\mathcal{D}$\\ ." ] }, { "cell_type": "raw", "metadata": { "raw_mimetype": "text/restructuredtext" }, "source": [ "\n", ".. admonition:: Example \\\n", "\n", " ✔ Consider PCA as a dimensionality reduction step, then :math:`f_j (x) = x {\\bf W}_j` where entries of the matrix :math:`{\\bf W}_j` are determined\n", " by the dataset :math:`{\\bf X} = [x_1 ; x_2 ; \\dots ; x_m]`\\ . Given a new data point :math:`\\bar{x}`\\ , the application of :math:`f_j` \n", " to :math:`\\bar{x}` is done by simple matrix multiplication.\n", " \\n\n", " \n", " ✘ Alternatively, consider the mapping :math:`f_j` obtained by fitting a PCoA (or `multidimensional scaling `_\\ ) \n", " with respect to some dissimilarity funtion. Then the mapping :math:`f_j` is relevant only to the samples that are in the training dataset. Applying it as a transformation \n", " to samples outside :math:`\\mathcal{D}` is not straight forward. \n", " \n" ] }, { "cell_type": "markdown", "metadata": { "raw_mimetype": "text/restructuredtext" }, "source": [ "In our terminology, functions that are easily applied to data outside of the training dataset are called **out-of-sample extendable**.\n", "\n", "\n", "## TCAM as dimensionality reduction step\n", "\n", "The TCAM is a $\\mathbb{R}^{n \\times p} \\to \\mathbb{R}^q$ mapping that is\n", "\n", "✔ Variance maximizing **(\\*)** \\\n", "✔ Distortion minimizing **(\\*)** \\\n", "✔ Out-of-sample extendable \n", "\n", "It makes perfect sense to use it as a dimensionality reduction step within ML workflows.\n", "\n", "\\ \\ \\ \\ **(\\*)** within the set of pseudo- $\\star_{\\bf M}$ orthogonal mappings of rank $q$ Mor et al.\n", "\n", "\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Concrete example\n", "\n", "We will now construct a concrete example for the incorporation of TCAM in supervised ML workflow for curve classification, using real world microbiome data.\n", "\n", "\n", "\n", "## Dataset\n", "\n", "The data in the following example was obtained from a study by Schirmer et al. Schirmer et al.\\ . In this study, the authors charactarized the gut microbiome throughout a disease course for 405 pediatric, new-onset, treatment-naive ulcerative colitis (UC) patients. \n", "\n", "\n", "Patients were monitored for 1 year upon treatment initiation, and microbial taxonomic composition was analyzed from fecal samples using 16S rRNA gene amplicon sequencing. \n", "Fecal samples were collected at baseline (week 0, prior to treatment) and 3 follow-up time points (4, 12, and 52 weeks after treatment initiation).\n", "At the end of the year, patients were assigned with \"label\" according their disease state, that is either \"Flare\" or \"Remission\".\n", "\n", "\n", "Formally: \\ \n", "\n", "* Inputs: $\\mathcal{X} \\subset \\mathbb{R}^{n \\times p}$ where $n$ are the four time points, and $p$ are the number of features obtained by 16S metagenomics sequencing. \n", "* Labels: $\\mathcal{Y} = \\{ \\operatorname{Flare}, \\operatorname{Remission} \\}$ \n", "\n", "\n", "We read the data from our git repository as follows" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Data\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Ruminococcaceae-g__Oscillospira-s__-OTU_310886k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Peptostreptococcaceae-g__-s__-OTU_531374k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__-g__-s__-OTU_366584k__Bacteria-p__Proteobacteria-c__Betaproteobacteria-o__Burkholderiales-f__Alcaligenaceae-g__Sutterella-s__-OTU_173726k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Ruminococcaceae-g__-s__-OTU_366794k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Lachnospiraceae-g__-s__-OTU_975306k__Bacteria-p__Bacteroidetes-c__Bacteroidia-o__Bacteroidales-f__Porphyromonadaceae-g__Parabacteroides-s__distasonis-OTU_585914k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__-g__-s__-OTU_352304k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Lachnospiraceae-g__Blautia-s__-OTU_193061k__Bacteria-p__Bacteroidetes-c__Bacteroidia-o__Bacteroidales-f__Prevotellaceae-g__Prevotella-s__-OTU_4302472...k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Lachnospiraceae-g__Roseburia-s__-OTU_313593k__Bacteria-p__Synergistetes-c__Synergistia-o__Synergistales-f__-g__-s__-OTU_505549k__Bacteria-p__Actinobacteria-c__Coriobacteriia-o__Coriobacteriales-f__Coriobacteriaceae-g__-s__-OTU_4403259k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Lachnospiraceae-g__Epulopiscium-s__-OTU_4371562k__Bacteria-p__Bacteroidetes-c__Bacteroidia-o__Bacteroidales-f__S24.7-g__-s__-OTU_1115121k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Ruminococcaceae-g__-s__-OTU_300870k__Bacteria-p__Bacteroidetes-c__Bacteroidia-o__Bacteroidales-f__Porphyromonadaceae-g__Parabacteroides-s__distasonis-OTU_186233k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Lachnospiraceae-g__.Ruminococcus-.s__-OTU_300877k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__.Tissierellaceae-.g__-s__-OTU_869796k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Ruminococcaceae-g__Ruminococcus-s__-OTU_196061
SubjectIDWeek
P_1034300.0000000.0000000.0000000.00.00.0008690.0229000.0000330.0000000.0...0.0000000.0000000.00.00.0000000.00.00.0000000.00.0
40.0001830.0000000.0089660.00.00.0204210.0000000.0008230.0000000.0...0.0000730.0000000.00.00.0000000.00.00.0000000.00.0
120.0000000.0000000.0000000.00.00.0062300.0000490.0009970.0000000.0...0.0000490.0000000.00.00.0000000.00.00.0000000.00.0
520.0000000.0000970.0000000.00.00.0022110.0000000.0002910.0000000.0...0.0000000.0000000.00.00.0000000.00.00.0000480.00.0
P_1089700.0004230.0000000.0000590.00.00.0772600.0008630.0174480.0000510.0...0.0000080.0000000.00.00.0000000.00.00.0000000.00.0
40.0000230.0000000.0000000.00.00.0060950.0004110.0000000.0000000.0...0.0000000.0000000.00.00.0000000.00.00.0000000.00.0
120.0000800.0000130.0108090.00.00.0054110.0003720.0003590.0000270.0...0.0000270.0000000.00.00.0000000.00.00.0000000.00.0
520.0000000.0000800.0006640.00.00.0139340.0002610.0009050.0000000.0...0.0000400.0000000.00.00.0000000.00.00.0000000.00.0
P_110800.0000000.0000000.0000000.00.00.0020860.0000000.0000660.0000160.0...0.0000490.0000000.00.00.0000490.00.00.0000000.00.0
40.0000000.0002110.0000000.00.00.0110070.0000000.0004090.0000000.0...0.0003840.0000250.00.00.0000000.00.00.0000000.00.0
120.0000070.0000000.0000280.00.00.0387600.0005280.0031050.0000000.0...0.0004280.0000000.00.00.0000000.00.00.0000000.00.0
520.0000000.0000000.0000000.00.00.1153250.0027980.0186510.0000000.0...0.0003110.0000000.00.00.0000000.00.00.0000000.00.0
\n", "

12 rows × 1015 columns

\n", "
" ], "text/plain": [ " k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Ruminococcaceae-g__Oscillospira-s__-OTU_310886 \\\n", "SubjectID Week \n", "P_10343 0 0.000000 \n", " 4 0.000183 \n", " 12 0.000000 \n", " 52 0.000000 \n", "P_10897 0 0.000423 \n", " 4 0.000023 \n", " 12 0.000080 \n", " 52 0.000000 \n", "P_1108 0 0.000000 \n", " 4 0.000000 \n", " 12 0.000007 \n", " 52 0.000000 \n", "\n", " k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Peptostreptococcaceae-g__-s__-OTU_531374 \\\n", "SubjectID Week \n", "P_10343 0 0.000000 \n", " 4 0.000000 \n", " 12 0.000000 \n", " 52 0.000097 \n", "P_10897 0 0.000000 \n", " 4 0.000000 \n", " 12 0.000013 \n", " 52 0.000080 \n", "P_1108 0 0.000000 \n", " 4 0.000211 \n", " 12 0.000000 \n", " 52 0.000000 \n", "\n", " k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__-g__-s__-OTU_366584 \\\n", "SubjectID Week \n", "P_10343 0 0.000000 \n", " 4 0.008966 \n", " 12 0.000000 \n", " 52 0.000000 \n", "P_10897 0 0.000059 \n", " 4 0.000000 \n", " 12 0.010809 \n", " 52 0.000664 \n", "P_1108 0 0.000000 \n", " 4 0.000000 \n", " 12 0.000028 \n", " 52 0.000000 \n", "\n", " k__Bacteria-p__Proteobacteria-c__Betaproteobacteria-o__Burkholderiales-f__Alcaligenaceae-g__Sutterella-s__-OTU_173726 \\\n", "SubjectID Week \n", "P_10343 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_10897 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_1108 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "\n", " k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Ruminococcaceae-g__-s__-OTU_366794 \\\n", "SubjectID Week \n", "P_10343 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_10897 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_1108 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "\n", " k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Lachnospiraceae-g__-s__-OTU_975306 \\\n", "SubjectID Week \n", "P_10343 0 0.000869 \n", " 4 0.020421 \n", " 12 0.006230 \n", " 52 0.002211 \n", "P_10897 0 0.077260 \n", " 4 0.006095 \n", " 12 0.005411 \n", " 52 0.013934 \n", "P_1108 0 0.002086 \n", " 4 0.011007 \n", " 12 0.038760 \n", " 52 0.115325 \n", "\n", " k__Bacteria-p__Bacteroidetes-c__Bacteroidia-o__Bacteroidales-f__Porphyromonadaceae-g__Parabacteroides-s__distasonis-OTU_585914 \\\n", "SubjectID Week \n", "P_10343 0 0.022900 \n", " 4 0.000000 \n", " 12 0.000049 \n", " 52 0.000000 \n", "P_10897 0 0.000863 \n", " 4 0.000411 \n", " 12 0.000372 \n", " 52 0.000261 \n", "P_1108 0 0.000000 \n", " 4 0.000000 \n", " 12 0.000528 \n", " 52 0.002798 \n", "\n", " k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__-g__-s__-OTU_352304 \\\n", "SubjectID Week \n", "P_10343 0 0.000033 \n", " 4 0.000823 \n", " 12 0.000997 \n", " 52 0.000291 \n", "P_10897 0 0.017448 \n", " 4 0.000000 \n", " 12 0.000359 \n", " 52 0.000905 \n", "P_1108 0 0.000066 \n", " 4 0.000409 \n", " 12 0.003105 \n", " 52 0.018651 \n", "\n", " k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Lachnospiraceae-g__Blautia-s__-OTU_193061 \\\n", "SubjectID Week \n", "P_10343 0 0.000000 \n", " 4 0.000000 \n", " 12 0.000000 \n", " 52 0.000000 \n", "P_10897 0 0.000051 \n", " 4 0.000000 \n", " 12 0.000027 \n", " 52 0.000000 \n", "P_1108 0 0.000016 \n", " 4 0.000000 \n", " 12 0.000000 \n", " 52 0.000000 \n", "\n", " k__Bacteria-p__Bacteroidetes-c__Bacteroidia-o__Bacteroidales-f__Prevotellaceae-g__Prevotella-s__-OTU_4302472 \\\n", "SubjectID Week \n", "P_10343 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_10897 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_1108 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "\n", " ... \\\n", "SubjectID Week ... \n", "P_10343 0 ... \n", " 4 ... \n", " 12 ... \n", " 52 ... \n", "P_10897 0 ... \n", " 4 ... \n", " 12 ... \n", " 52 ... \n", "P_1108 0 ... \n", " 4 ... \n", " 12 ... \n", " 52 ... \n", "\n", " k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Lachnospiraceae-g__Roseburia-s__-OTU_313593 \\\n", "SubjectID Week \n", "P_10343 0 0.000000 \n", " 4 0.000073 \n", " 12 0.000049 \n", " 52 0.000000 \n", "P_10897 0 0.000008 \n", " 4 0.000000 \n", " 12 0.000027 \n", " 52 0.000040 \n", "P_1108 0 0.000049 \n", " 4 0.000384 \n", " 12 0.000428 \n", " 52 0.000311 \n", "\n", " k__Bacteria-p__Synergistetes-c__Synergistia-o__Synergistales-f__-g__-s__-OTU_505549 \\\n", "SubjectID Week \n", "P_10343 0 0.000000 \n", " 4 0.000000 \n", " 12 0.000000 \n", " 52 0.000000 \n", "P_10897 0 0.000000 \n", " 4 0.000000 \n", " 12 0.000000 \n", " 52 0.000000 \n", "P_1108 0 0.000000 \n", " 4 0.000025 \n", " 12 0.000000 \n", " 52 0.000000 \n", "\n", " k__Bacteria-p__Actinobacteria-c__Coriobacteriia-o__Coriobacteriales-f__Coriobacteriaceae-g__-s__-OTU_4403259 \\\n", "SubjectID Week \n", "P_10343 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_10897 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_1108 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "\n", " k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Lachnospiraceae-g__Epulopiscium-s__-OTU_4371562 \\\n", "SubjectID Week \n", "P_10343 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_10897 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_1108 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "\n", " k__Bacteria-p__Bacteroidetes-c__Bacteroidia-o__Bacteroidales-f__S24.7-g__-s__-OTU_1115121 \\\n", "SubjectID Week \n", "P_10343 0 0.000000 \n", " 4 0.000000 \n", " 12 0.000000 \n", " 52 0.000000 \n", "P_10897 0 0.000000 \n", " 4 0.000000 \n", " 12 0.000000 \n", " 52 0.000000 \n", "P_1108 0 0.000049 \n", " 4 0.000000 \n", " 12 0.000000 \n", " 52 0.000000 \n", "\n", " k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Ruminococcaceae-g__-s__-OTU_300870 \\\n", "SubjectID Week \n", "P_10343 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_10897 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_1108 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "\n", " k__Bacteria-p__Bacteroidetes-c__Bacteroidia-o__Bacteroidales-f__Porphyromonadaceae-g__Parabacteroides-s__distasonis-OTU_186233 \\\n", "SubjectID Week \n", "P_10343 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_10897 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_1108 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "\n", " k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Lachnospiraceae-g__.Ruminococcus-.s__-OTU_300877 \\\n", "SubjectID Week \n", "P_10343 0 0.000000 \n", " 4 0.000000 \n", " 12 0.000000 \n", " 52 0.000048 \n", "P_10897 0 0.000000 \n", " 4 0.000000 \n", " 12 0.000000 \n", " 52 0.000000 \n", "P_1108 0 0.000000 \n", " 4 0.000000 \n", " 12 0.000000 \n", " 52 0.000000 \n", "\n", " k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__.Tissierellaceae-.g__-s__-OTU_869796 \\\n", "SubjectID Week \n", "P_10343 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_10897 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_1108 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "\n", " k__Bacteria-p__Firmicutes-c__Clostridia-o__Clostridiales-f__Ruminococcaceae-g__Ruminococcus-s__-OTU_196061 \n", "SubjectID Week \n", "P_10343 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_10897 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "P_1108 0 0.0 \n", " 4 0.0 \n", " 12 0.0 \n", " 52 0.0 \n", "\n", "[12 rows x 1015 columns]" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Labels\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
SubjectIDlabel
0P_103430
1P_103430
2P_103430
3P_103430
\n", "
" ], "text/plain": [ " SubjectID label\n", "0 P_10343 0\n", "1 P_10343 0\n", "2 P_10343 0\n", "3 P_10343 0" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import pandas as pd\n", "import numpy as np\n", "\n", "data_raw = pd.read_csv(\"https://raw.githubusercontent.com/UriaMorP/\"\n", " \"tcam_analysis_notebooks/main/Schirmer2018/Schirmer2018.tsv\"\n", " , index_col=[0,1], sep=\"\\t\"\n", " , dtype={'Week':int})\n", "\n", "\n", "labels = pd.read_csv(\"https://raw.githubusercontent.com/UriaMorP/\"\n", " \"tcam_analysis_notebooks/main/Schirmer2018/metadata_Schirmer2018.tsv\"\n", " , index_col=[0], sep=\"\\t\").loc[data_raw.index.get_level_values(0), 'Remission'].rename('label')\n", "labels = 1 * (labels == \"Remission\") # Binarize the labels\n", "labels = labels.reset_index().copy()\n", "\n", "\n", "print(\"Data\")\n", "display(data_raw.head(12))\n", "print(\"Labels\")\n", "display(labels.head(4))\n", "\n", "\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Making the ML workflow\n", "\n", "We rely on sklearn's ``Pipeline`` interface in order to implement the composition of functions, previously formulated as $f = f_1 \\circ f_2 \\circ ... \\circ f_d$ making the classification model. \n", "\n", "A python code realization of the above mathematical formulation would be as follows:\n", "\n", "```python\n", "from sklearn.pipeline import Pipeline\n", "\n", "pipe = Pipeline([\n", " ('f_1', Transformer1())\n", " ,('f_2', Transformer2()) \n", " # ...\n", " ,('f_d', Estimator()) \n", "])\n", "\n", "```\n", "\n", "where ``Transformer``s are classes implementing ``sklearn.base.TransformerMixin`` and ``Estimator`` implements ``sklearn.base.BaseEstimator``\\ .\n", "\n", "\n", "In this example, we use the following steps \\\n", "\n", "1. Filtration of low-abundant features \\\n", "2. *Curvewise* standardization \\\n", "\n", "3. TCAM truncation \\\n", "4. MinMax scaling \\\n", "5. GradientBoostClassifier \\\n", "\n", "\n", "For steps 1.-3. it is most convenient to work with ``pandas.DataFrame``s as inputs. " ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [], "source": [ "from sklearn.base import BaseEstimator, TransformerMixin\n", "\n", "# step 1: Filtration of low-abundant features\n", "class RAFilter(BaseEstimator, TransformerMixin):\n", " def __init__(self, capval = 5e-2):\n", " self._f_list = []\n", " self.capval = capval\n", " pass\n", "\n", " def fit(self, X,y=None, modemap = []):\n", " X_t = X.copy()\n", " X_t = 100 * (X_t / X_t.values.sum(axis=1,keepdims = True))\n", " self._f_list = [i for i in range(X_t.shape[1]) if (X_t.iloc[:,i] > self.capval).sum() > X_t.shape[0]*0.1 ]\n", " return self\n", " \n", " def transform(self, X, y = None):\n", " X_t = 100 * (X / X.values.sum(axis=1,keepdims = True)).copy()\n", " X_t = X_t.iloc[:,self._f_list]\n", " X_t[X_t < self.capval] = self.capval\n", " X_t = 100 * X_t / X_t.values.sum(axis=1,keepdims = True)\n", " return X_t\n", "\n", " def fit_transform(self, X, y = None, modemap = []):\n", " self = self.fit(X, y)\n", " Xt = self.transform(X, y)\n", " return Xt" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Since microbiome data tends to cluster by participant, meaning that variation coming from interindividual differences tend overshadow any other source of variation, including temporal trends, the task of identifying mutual temporal trajectories in microbiome composition is very challenging.
\n", "Keeping in mind that we are only concerned with the direction of temporal variation trends (rather than the concrete baseline), one simple way of tackling this challenge would be to look at deviations from the baseline period.
\n", "That is, instead of looking at raw values $x_{ij}$ (the $j^{th}$ time point of subject $i$), we consider tthe followoing:\n", "\n", "$$\n", "z_{ij} = \\log_{2}(x_{ij} / \\mu(x_{i0}) )\n", "$$\n", "\n", "where $\\mu ( x_{i0} )$ denotes the baseline sample of subject $i$ . Using pandas, this transfromation looks like:" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [], "source": [ "from sklearn.preprocessing import FunctionTransformer\n", "\n", "# step 2: Curvewise standardization\n", "def standardize_curve_df(X, y = None, cap = 5e-3):\n", " blmean = (X + cap).groupby(level=['SubjectID']).apply(lambda x:x / x.loc[x.index.get_level_values('Week') == 0].iloc[0] )\n", " blmean = np.log2(blmean)\n", " return blmean\n", "\n", "# Wrap the function with Transformer\n", "StandardCurve = FunctionTransformer(standardize_curve_df) " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Step 3. is the last step applied to a dataframe. \n", "In this step, we cast the dataframe into a tensor, then a TCAM is applied, followed by truncation (dimensionality reduction step) and a ``numpy.array`` is returned. \n", "\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "
\n", " \n", "\n", "Gotcha\n", "\n", "Suppose that the size of the training dataset is $4m$ in its tabular form, the number of rows in the output of ``TCAMWrapper.transform`` is $m$\n", "\n", "For many good reasons, this reduction in number of **samples** , is not allowed by the ``Pipeline`` (it throws a `ValueError: Found input variables with inconsistent numbers of samples: [m, 4m]` )\n", " \n", "In our case, this decrease is intentional, and we hack our way around sklearn's safeguards\n", "\n", "
\n", "\n", "To overcome this inconsistency, we divide the workflow into two pipelines each of which maintains consistent number of rows, and the TCAM transformation is patched between the two pipelines." ] }, { "cell_type": "raw", "metadata": { "raw_mimetype": "text/restructuredtext" }, "source": [ ".. figure:: ./data/mermaid-diagram-pipeline.png\n", " :alt: memaid-pipes\n", " :class: with-shadow\n", " :width: 50%\n", " :align: center\n", "\n", " The pipeline patching process, the inputs and outputs of ``Pipeline1`` all have :math:`4m` rows, while those of ``Pipeline1`` have :math:`m` rows. \n", " The ``TCAM`` transformation (of matrix to tensors) takes inputs with :math:`4m` rows and outputs tables of :math:`m` rows. " ] }, { "cell_type": "code", "execution_count": 118, "metadata": {}, "outputs": [], "source": [ "from sklearn.pipeline import Pipeline\n", "from sklearn.preprocessing import MinMaxScaler\n", "from sklearn.neural_network import MLPClassifier\n", "from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier\n", "from mprod.dimensionality_reduction import TCAM\n", "from mprod import table2tensor\n", "\n", "class Patch(BaseEstimator, TransformerMixin):\n", " def __init__(self,pipeline1, tcam_obj):\n", " self.pipeline1 = pipeline1\n", " self.tcam_obj = tcam_obj\n", " \n", " def _transform_y(self, y, map1):\n", " yt = pd.Series(map1).rename('row').to_frame()\n", " yt = yt.merge(y.reset_index(), left_index = True, right_on = \"SubjectID\").drop_duplicates()\n", " yt.index = yt['row']\n", " yt = yt['label'].sort_index()\n", " return yt\n", "\n", " \n", " def fit(self, X, y = None):\n", " self.pipeline1 = self.pipeline1.fit(X,y=y)\n", " tensor, map1, map3 = table2tensor(self.pipeline1.transform(X))\n", " \n", " if y is not None:\n", " yt = self._transform_y(y,map1)\n", " else:\n", " yt = None\n", " \n", " self.tcam_obj = self.tcam_obj.fit(tensor,y=yt)\n", " return self\n", " \n", " def transform(self, X):\n", " Xt1 = self.pipeline1.transform(X)\n", " tensor, map1, map3 = table2tensor(Xt1)\n", " Xt2 = self.tcam_obj.transform(tensor)\n", " return Xt2" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Instantiation on this workflow is done as follows:" ] }, { "cell_type": "code", "execution_count": 119, "metadata": {}, "outputs": [], "source": [ "rs = 0 # To make sure executions are consistant \n", "pipeline1 = Pipeline([\n", " (\"raf\", RAFilter(capval = 5e-3)),\n", " (\"st_curve\", StandardCurve)\n", "])\n", "\n", "pipeline2 = Pipeline([\n", " ('mmt',MinMaxScaler(clip = True)),\n", " ('clf', GradientBoostingClassifier(learning_rate=.2,random_state=rs))\n", "])\n", "\n", "patch = Pipeline([\n", " ('pipeline1',Patch(pipeline1, TCAM(n_components=.8)))\n", " , ('pipeline2', pipeline2)])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Fron here and on, the ``patch`` object is just like any other classification model, and can be evaluated as such\n", "\n", "\n", "### Model evaluation" ] }, { "cell_type": "code", "execution_count": 120, "metadata": { "scrolled": false }, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "from sklearn.model_selection import StratifiedKFold\n", "from sklearn.metrics import auc\n", "from sklearn.metrics import plot_roc_curve, roc_curve\n", "import matplotlib.pyplot as plt\n", "%matplotlib inline\n", "\n", "cv = StratifiedKFold(n_splits=6, shuffle=True, random_state=rs)\n", "\n", "# indices for CV iterations\n", "sid_ser = labels.copy().drop_duplicates()\n", "sid_ser.index = sid_ser['SubjectID']\n", "sid_ser = sid_ser['label']\n", "dummy_size = np.arange(sid_ser.values.size)\n", "\n", "# storing the model's scores\n", "tprs = []\n", "aucs = []\n", "mean_fpr = np.linspace(0, 1, 100)\n", "\n", "# CV loop\n", "fig, ax = plt.subplots()\n", "for i, (train_iloc, test_iloc) in enumerate(cv.split(dummy_size,sid_ser.values, sid_ser.values)):\n", " \n", " train_id = sid_ser.index[train_iloc].copy().sort_values()\n", " test_id = sid_ser.index[test_iloc].copy().sort_values()\n", "\n", " y_train = sid_ser[train_id].copy()\n", " X_train = data_raw.loc[data_raw.index.get_level_values(\"SubjectID\").isin(y_train.index)].copy().sort_index(0)\n", " \n", " y_test = sid_ser[test_id].copy()\n", " X_test = data_raw.loc[data_raw.index.get_level_values(\"SubjectID\").isin(y_test.index)].copy().sort_index(0)\n", " \n", " trained_model = patch.fit(X_train, y_train.copy())\n", " viz = plot_roc_curve(trained_model, X_test, y_test.copy(),\n", " name='F{}'.format(i+1),\n", " alpha=0.3, lw=1, ax=ax)\n", " \n", " interp_tpr = np.interp(mean_fpr, viz.fpr, viz.tpr)\n", " interp_tpr[0] = 0.0\n", "\n", " tprs.append(interp_tpr)\n", " aucs.append(viz.roc_auc)\n", "\n", "ax.plot([0, 1], [0, 1], linestyle=\"--\", lw=2, color=\"r\", label=\"Chance\", alpha=0.8)\n", "\n", "mean_tpr = np.mean(tprs, axis=0)\n", "mean_tpr[-1] = 1.0\n", "mean_auc = auc(mean_fpr, mean_tpr)\n", "std_auc = np.std(aucs)\n", "ax.plot(\n", " mean_fpr,\n", " mean_tpr,\n", " color=\"b\",\n", " label=r\"Mean AUROC = %0.2f \" % (mean_auc),\n", " lw=2,\n", " alpha=0.8,\n", ")\n", "\n", "std_tpr = np.std(tprs, axis=0)\n", "tprs_upper = np.minimum(mean_tpr + std_tpr, 1)\n", "tprs_lower = np.maximum(mean_tpr - std_tpr, 0)\n", "ax.fill_between(\n", " mean_fpr,\n", " tprs_lower,\n", " tprs_upper,\n", " color=\"grey\",\n", " alpha=0.2,\n", " label=r\"$\\pm$ 1 std. dev.\",\n", ")\n", "\n", "ax.set(\n", " xlim=[-0.05, 1.05],\n", " ylim=[-0.05, 1.05],\n", ")\n", "ax.legend(loc=[.6, 0.05], fontsize = 6 )\n", "ax.set_ylabel(\"TPR\")\n", "ax.set_xlabel(\"FPR\\n(positive label: 1)\")\n", "fig.set_size_inches([2.1,2.1])\n", "fig.set_dpi(300)\n", "plt.show()\n", "\n", " \n", " \n", " \n" ] }, { "cell_type": "raw", "metadata": { "raw_mimetype": "text/restructuredtext" }, "source": [ ".. footbibliography::" ] } ], "metadata": { "celltoolbar": "Raw Cell Format", "kernelspec": { "display_name": "mprod", "language": "python", "name": "mprod" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.6.8" } }, "nbformat": 4, "nbformat_minor": 4 }