{ "cells": [ { "cell_type": "markdown", "id": "05d93248", "metadata": {}, "source": [ "" ] }, { "cell_type": "markdown", "id": "4744f004", "metadata": {}, "source": [ "# Multi-Class Categorical Comparisons" ] }, { "cell_type": "code", "execution_count": 1, "id": "275a7087", "metadata": { "tags": [ "hide-output" ] }, "outputs": [], "source": [ "import rioxarray as rxr\n", "import gval\n", "import numpy as np\n", "import pandas as pd\n", "import xarray as xr\n", "from itertools import product\n", "\n", "pd.set_option('display.max_columns', None)" ] }, { "cell_type": "markdown", "id": "34069943", "metadata": {}, "source": [ "## Load Datasets" ] }, { "cell_type": "code", "execution_count": 2, "id": "38473c06", "metadata": {}, "outputs": [], "source": [ "candidate = rxr.open_rasterio(\n", " \"./candidate_map_multi_categorical.tif\", mask_and_scale=True\n", ")\n", "benchmark = rxr.open_rasterio(\n", " \"./benchmark_map_multi_categorical.tif\", mask_and_scale=True\n", ")\n", "depth_raster = rxr.open_rasterio(\n", " \"./candidate_raw_elevation_multi_categorical.tif\", mask_and_scale=True\n", ")" ] }, { "cell_type": "markdown", "id": "fa522035", "metadata": {}, "source": [ "## Homogenize Datasets and Make Agreement Map" ] }, { "cell_type": "markdown", "id": "e3e5ca15", "metadata": {}, "source": [ "Although one can call `candidate.gval.categorical_compare` to run the entire workflow, in this case homogenization and creation of an agreement map will be done separately to show more options for multi-class comparisons." ] }, { "cell_type": "markdown", "id": "2ac66a26", "metadata": {}, "source": [ "#### Homogenize" ] }, { "cell_type": "code", "execution_count": 3, "id": "29375e17", "metadata": {}, "outputs": [], "source": [ "candidate_r, benchmark_r = candidate.gval.homogenize(benchmark)\n", "depth_raster_r, arb = depth_raster.gval.homogenize(benchmark_r)\n", "del arb" ] }, { "cell_type": "markdown", "id": "4e9e1be1", "metadata": {}, "source": [ "#### Agreement Map" ] }, { "cell_type": "markdown", "id": "e2851c9b", "metadata": {}, "source": [ "The following makes a pairing dictionary which maps combinations of values in the candidate and benchmark maps to unique values in the agreement map. In this case we will encode each value as concatenation of what the values are. Instead of making a pairing dictionary one can use the `szudzik` or `cantor` pairing functions to make unique values for each combination of candidate and benchmark map values. e.g. 12 represents a class 1 for the candidate and a class 2 for the benchmark." ] }, { "cell_type": "code", "execution_count": 4, "id": "de894568", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "(1, 1): 11\n", "(1, 2): 12\n", "(1, 3): 13\n", "(1, 4): 14\n", "(1, 5): 15\n", "(2, 1): 21\n" ] } ], "source": [ "classes = [1, 2, 3, 4, 5]\n", "pairing_dictionary = {(x, y): int(f'{x}{y}') for x, y in product(*([classes]*2))}\n", "\n", "# Showing the first 6 entries\n", "print('\\n'.join([f'{k}: {v}' for k,v in pairing_dictionary.items()][:6]))" ] }, { "cell_type": "markdown", "id": "44328dcf", "metadata": {}, "source": [ "The benchmark map has an extra class 0 which is very similar to nodata so it will not be included in `allow_benchmark_values` in the following methods." ] }, { "cell_type": "code", "execution_count": 5, "id": "1dc16dd7", "metadata": {}, "outputs": [], "source": [ "agreement_map = candidate_r.gval.compute_agreement_map(\n", " benchmark_r,\n", " nodata=255,\n", " encode_nodata=True,\n", " comparison_function=\"pairing_dict\",\n", " pairing_dict=pairing_dictionary,\n", " allow_candidate_values=classes,\n", " allow_benchmark_values=classes,\n", ")\n", "\n", "crosstab = agreement_map.gval.compute_crosstab()" ] }, { "cell_type": "markdown", "id": "93fe86df", "metadata": {}, "source": [ "The following only shows a small subset of the map for memory purposes:" ] }, { "cell_type": "code", "execution_count": 7, "id": "55606165", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "agreement_map.gval.cat_plot(\n", " title='Agreement Map', \n", " figsize=(8, 6),\n", " colormap='tab20b'\n", ")" ] }, { "cell_type": "markdown", "id": "bdcbfb8e", "metadata": {}, "source": [ "## Comparisons" ] }, { "cell_type": "markdown", "id": "4a1f3ecc", "metadata": {}, "source": [ "For multi-categorical statistics GVAL offers 4 methods of averaging:\n", "\n", "1. No Averaging which provides one vs. all metrics on a class basis\n", "1. Micro Averaging which sums up the contingencies of each class defined as either positive or negative\n", "3. Macro Averaging which sums up the contingencies of one class vs all and then averages them\n", "4. Weighted Averaging which does macro averaging with the inclusion of weights to be applied to each positive category." ] }, { "cell_type": "markdown", "id": "66235a0a", "metadata": {}, "source": [ "### No Averaging" ] }, { "cell_type": "markdown", "id": "4f258087", "metadata": {}, "source": [ "Using `None` for the averaging argument runs a one class vs. all methodology for each class and reports their metrics on a class basis:" ] }, { "cell_type": "code", "execution_count": 8, "id": "936f2dea", "metadata": {}, "outputs": [ { "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", "
01234
band11111
positive_categories12345
fn6.01043.0318274.0516572.0364147.0
fp172762.0561004.0462496.03775.05.0
tn1043592.0653360.0422623.0693617.0852206.0
tp0.0953.012967.02396.02.0
accuracy0.8579630.5379270.3581090.572210.700622
balanced_accuracy0.4289840.5077410.2583110.4996020.5
critical_success_index0.00.0016930.0163370.0045840.000005
equitable_threat_score-0.0000050.000055-0.175401-0.000455-0.0
f_score0.00.003380.0321480.0091250.000011
false_discovery_rate1.00.9983040.9727280.6117320.714286
false_negative_rate1.00.5225450.9608530.9953830.999995
false_omission_rate0.0000060.0015940.4295790.4268520.299376
false_positive_rate0.1420330.4619740.5225240.0054130.000006
fowlkes_mallows_index0.00.0284550.0326750.0423390.001253
matthews_correlation_coefficient-0.0009040.001257-0.440983-0.005543-0.000072
negative_likelihood_ratio1.1655460.9712262.012361.0008011.0
negative_predictive_value0.9999940.9984060.5704210.5731480.700624
overall_bias28793.666667281.5415831.4353990.0118910.000019
positive_likelihood_ratio0.01.0335110.0749190.8529160.936112
positive_predictive_value0.00.0016960.0272720.3882680.285714
prevalence0.0000050.0016410.2723220.4266570.299376
prevalence_threshold1.00.495880.7851070.5198760.508252
true_negative_rate0.8579670.5380260.4774760.9945870.999994
true_positive_rate0.00.4774550.0391470.0046170.000005
\n", "
" ], "text/plain": [ " 0 1 2 \\\n", "band 1 1 1 \n", "positive_categories 1 2 3 \n", "fn 6.0 1043.0 318274.0 \n", "fp 172762.0 561004.0 462496.0 \n", "tn 1043592.0 653360.0 422623.0 \n", "tp 0.0 953.0 12967.0 \n", "accuracy 0.857963 0.537927 0.358109 \n", "balanced_accuracy 0.428984 0.507741 0.258311 \n", "critical_success_index 0.0 0.001693 0.016337 \n", "equitable_threat_score -0.000005 0.000055 -0.175401 \n", "f_score 0.0 0.00338 0.032148 \n", "false_discovery_rate 1.0 0.998304 0.972728 \n", "false_negative_rate 1.0 0.522545 0.960853 \n", "false_omission_rate 0.000006 0.001594 0.429579 \n", "false_positive_rate 0.142033 0.461974 0.522524 \n", "fowlkes_mallows_index 0.0 0.028455 0.032675 \n", "matthews_correlation_coefficient -0.000904 0.001257 -0.440983 \n", "negative_likelihood_ratio 1.165546 0.971226 2.01236 \n", "negative_predictive_value 0.999994 0.998406 0.570421 \n", "overall_bias 28793.666667 281.541583 1.435399 \n", "positive_likelihood_ratio 0.0 1.033511 0.074919 \n", "positive_predictive_value 0.0 0.001696 0.027272 \n", "prevalence 0.000005 0.001641 0.272322 \n", "prevalence_threshold 1.0 0.49588 0.785107 \n", "true_negative_rate 0.857967 0.538026 0.477476 \n", "true_positive_rate 0.0 0.477455 0.039147 \n", "\n", " 3 4 \n", "band 1 1 \n", "positive_categories 4 5 \n", "fn 516572.0 364147.0 \n", "fp 3775.0 5.0 \n", "tn 693617.0 852206.0 \n", "tp 2396.0 2.0 \n", "accuracy 0.57221 0.700622 \n", "balanced_accuracy 0.499602 0.5 \n", "critical_success_index 0.004584 0.000005 \n", "equitable_threat_score -0.000455 -0.0 \n", "f_score 0.009125 0.000011 \n", "false_discovery_rate 0.611732 0.714286 \n", "false_negative_rate 0.995383 0.999995 \n", "false_omission_rate 0.426852 0.299376 \n", "false_positive_rate 0.005413 0.000006 \n", "fowlkes_mallows_index 0.042339 0.001253 \n", "matthews_correlation_coefficient -0.005543 -0.000072 \n", "negative_likelihood_ratio 1.000801 1.0 \n", "negative_predictive_value 0.573148 0.700624 \n", "overall_bias 0.011891 0.000019 \n", "positive_likelihood_ratio 0.852916 0.936112 \n", "positive_predictive_value 0.388268 0.285714 \n", "prevalence 0.426657 0.299376 \n", "prevalence_threshold 0.519876 0.508252 \n", "true_negative_rate 0.994587 0.999994 \n", "true_positive_rate 0.004617 0.000005 " ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" } ], "source": [ "no_averaged_metrics = crosstab.gval.compute_categorical_metrics(\n", " positive_categories=[1, 2, 3, 4, 5],\n", " negative_categories=None,\n", " average=None\n", ")\n", "no_averaged_metrics.transpose()" ] }, { "cell_type": "markdown", "id": "d722dc68", "metadata": {}, "source": [ "### Micro Averaging" ] }, { "cell_type": "markdown", "id": "3bbb83cf", "metadata": {}, "source": [ "The following is an example of a using micro averaging to combine classes to process two-class categorical statistics. In this example we will use classes 1 and 2 as positive classes and classes 3, 4, and 5 as negative classes:" ] }, { "cell_type": "code", "execution_count": 9, "id": "538dfc49", "metadata": {}, "outputs": [ { "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", "
0
band1
fn382.0
fp733099.0
tn481259.0
tp1620.0
accuracy0.396987
balanced_accuracy0.602749
critical_success_index0.002204
equitable_threat_score0.00056
f_score0.004398
false_discovery_rate0.997795
false_negative_rate0.190809
false_omission_rate0.000793
false_positive_rate0.603693
fowlkes_mallows_index0.04224
matthews_correlation_coefficient0.017033
negative_likelihood_ratio0.481468
negative_predictive_value0.999207
overall_bias366.992507
positive_likelihood_ratio1.340402
positive_predictive_value0.002205
prevalence0.001646
prevalence_threshold0.463444
true_negative_rate0.396307
true_positive_rate0.809191
\n", "
" ], "text/plain": [ " 0\n", "band 1\n", "fn 382.0\n", "fp 733099.0\n", "tn 481259.0\n", "tp 1620.0\n", "accuracy 0.396987\n", "balanced_accuracy 0.602749\n", "critical_success_index 0.002204\n", "equitable_threat_score 0.00056\n", "f_score 0.004398\n", "false_discovery_rate 0.997795\n", "false_negative_rate 0.190809\n", "false_omission_rate 0.000793\n", "false_positive_rate 0.603693\n", "fowlkes_mallows_index 0.04224\n", "matthews_correlation_coefficient 0.017033\n", "negative_likelihood_ratio 0.481468\n", "negative_predictive_value 0.999207\n", "overall_bias 366.992507\n", "positive_likelihood_ratio 1.340402\n", "positive_predictive_value 0.002205\n", "prevalence 0.001646\n", "prevalence_threshold 0.463444\n", "true_negative_rate 0.396307\n", "true_positive_rate 0.809191" ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" } ], "source": [ "micro_averaged_metrics = crosstab.gval.compute_categorical_metrics(\n", " positive_categories=[1, 2],\n", " negative_categories=[3, 4, 5],\n", " average=\"micro\"\n", ")\n", "micro_averaged_metrics.transpose()" ] }, { "cell_type": "markdown", "id": "79761a73", "metadata": {}, "source": [ "### Macro Averaging" ] }, { "cell_type": "markdown", "id": "790c56df", "metadata": {}, "source": [ "The following shows macro averaging and is equivalent to the values of shared columns in `no_averaged_comps.mean()`:" ] }, { "cell_type": "code", "execution_count": 10, "id": "7e64eb9b", "metadata": {}, "outputs": [ { "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", "
0
band1
accuracy0.605366
balanced_accuracy0.438927
critical_success_index0.004524
equitable_threat_score-0.035161
f_score0.008933
false_discovery_rate0.85941
false_negative_rate0.895755
false_omission_rate0.231481
false_positive_rate0.22639
fowlkes_mallows_index0.020944
matthews_correlation_coefficient-0.089249
negative_likelihood_ratio1.229986
negative_predictive_value0.768519
overall_bias5815.331112
positive_likelihood_ratio0.579492
positive_predictive_value0.14059
prevalence0.2
prevalence_threshold0.661823
true_negative_rate0.77361
true_positive_rate0.104245
\n", "
" ], "text/plain": [ " 0\n", "band 1\n", "accuracy 0.605366\n", "balanced_accuracy 0.438927\n", "critical_success_index 0.004524\n", "equitable_threat_score -0.035161\n", "f_score 0.008933\n", "false_discovery_rate 0.85941\n", "false_negative_rate 0.895755\n", "false_omission_rate 0.231481\n", "false_positive_rate 0.22639\n", "fowlkes_mallows_index 0.020944\n", "matthews_correlation_coefficient -0.089249\n", "negative_likelihood_ratio 1.229986\n", "negative_predictive_value 0.768519\n", "overall_bias 5815.331112\n", "positive_likelihood_ratio 0.579492\n", "positive_predictive_value 0.14059\n", "prevalence 0.2\n", "prevalence_threshold 0.661823\n", "true_negative_rate 0.77361\n", "true_positive_rate 0.104245" ] }, "execution_count": 10, "metadata": {}, "output_type": "execute_result" } ], "source": [ "macro_averaged_metrics = crosstab.gval.compute_categorical_metrics(\n", " positive_categories=classes,\n", " negative_categories=None,\n", " average=\"macro\"\n", ")\n", "macro_averaged_metrics.transpose()" ] }, { "cell_type": "markdown", "id": "ef8f72ab", "metadata": {}, "source": [ "### Weighted Averaging" ] }, { "cell_type": "markdown", "id": "e182a6f7", "metadata": {}, "source": [ "To further enhance `macro-averaging`, we can apply weights to the classes of interest in order to appropriately change the strength of evaluations for each class. For instance, if we applied the following vector the classes uses in this notebook, `[1, 4, 1, 5, 1]`, classes 2 and 4 would have greater influence on the final averaging of the scores for all classes. (All weight values are in reference to the other weight values respectively. e.g. the vector `[5, 5, 5, 5, 5]` would cause no change in the averaging because each weight value is equivalent to a ll other weight values.) Let's use the first weight vector mentioned in weighted averaging:" ] }, { "cell_type": "code", "execution_count": 11, "id": "0eae1cbc", "metadata": {}, "outputs": [ { "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", "
0
band1
accuracy0.577454
balanced_accuracy0.476356
critical_success_index0.003836
equitable_threat_score-0.014789
f_score0.007609
false_discovery_rate0.811574
false_negative_rate0.835662
false_omission_rate0.239133
false_positive_rate0.211627
fowlkes_mallows_index0.029953
matthews_correlation_coefficient-0.03872
negative_likelihood_ratio1.088901
negative_predictive_value0.760867
overall_bias2493.443989
positive_likelihood_ratio0.784138
positive_predictive_value0.188426
prevalence0.225962
prevalence_threshold0.573022
true_negative_rate0.788373
true_positive_rate0.164338
\n", "
" ], "text/plain": [ " 0\n", "band 1\n", "accuracy 0.577454\n", "balanced_accuracy 0.476356\n", "critical_success_index 0.003836\n", "equitable_threat_score -0.014789\n", "f_score 0.007609\n", "false_discovery_rate 0.811574\n", "false_negative_rate 0.835662\n", "false_omission_rate 0.239133\n", "false_positive_rate 0.211627\n", "fowlkes_mallows_index 0.029953\n", "matthews_correlation_coefficient -0.03872\n", "negative_likelihood_ratio 1.088901\n", "negative_predictive_value 0.760867\n", "overall_bias 2493.443989\n", "positive_likelihood_ratio 0.784138\n", "positive_predictive_value 0.188426\n", "prevalence 0.225962\n", "prevalence_threshold 0.573022\n", "true_negative_rate 0.788373\n", "true_positive_rate 0.164338" ] }, "execution_count": 11, "metadata": {}, "output_type": "execute_result" } ], "source": [ "weight_averaged_metrics = crosstab.gval.compute_categorical_metrics(\n", " positive_categories=classes,\n", " weights=[1, 4, 1, 5, 1],\n", " negative_categories=None,\n", " average=\"weighted\"\n", ")\n", "weight_averaged_metrics.transpose()" ] }, { "cell_type": "markdown", "id": "8c567b77", "metadata": {}, "source": [ "Regardless of the averaging methodology it seems as though the candidate does not agree with the benchmark. We can now save the output." ] }, { "cell_type": "markdown", "id": "0d5f7be8", "metadata": {}, "source": [ "## Save Output" ] }, { "cell_type": "code", "execution_count": 12, "id": "dff8f8a0", "metadata": {}, "outputs": [], "source": [ "# output agreement map\n", "agreement_file = 'multi_categorical_agreement_map.tif'\n", "metric_file = 'macro_averaged_metric_file.csv'\n", "\n", "agreement_map.rio.to_raster(agreement_file)\n", "macro_averaged_metrics.to_csv(metric_file)" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.10.10" } }, "nbformat": 4, "nbformat_minor": 5 }