42_random_hier_rand.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "8816b366-5be4-4748-b769-4dc6f7d1735f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Modify the file 00_setup.py to define input/output file paths on your system\n",
    "# The information in 00_setup.py will be used across notebooks\n",
    "from importlib.machinery import SourceFileLoader\n",
    "setup = SourceFileLoader(\"setup\", \"./00_setup.py\").load_module()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a945de0c-f042-46b2-9d06-ab064af6c05b",
   "metadata": {},
   "source": [
    "## Random Hierarchy for NAICS (fixed randomization modifications)\n",
    "Neural Network With NAICS Embedding Layer + dummy hieararchy. Add NAICS embedding to model.  Add additional dummy levels to the hieararchy.   Modify training data to inject 10% 1 values for all levels"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e169c2e8-ba0f-488f-acf6-87d419bd4e99",
   "metadata": {},
   "source": [
    "Use the dataset with no missing values. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d47a7ec9-7d7b-47df-a902-b2f1e54f1135",
   "metadata": {},
   "source": [
    "*This script takes about 2 hours on my MacBook Air*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "fb42f9b4-99a8-4753-aacf-8722469be4c8",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "bf51170a-4487-4d10-b63d-c88f41b529f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pathlib import Path\n",
    "import importlib, pickle\n",
    "from sklearn import model_selection\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "09dd973b-60a0-434b-9193-8c5fcec27dd5",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import os\n",
    "import re\n",
    "\n",
    "import keras\n",
    "from tensorflow.keras import layers, optimizers, losses, metrics, Model\n",
    "from sklearn import preprocessing, feature_extraction, model_selection\n",
    "from IPython.display import display, HTML\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "22b466e7-a419-45aa-83f7-5dafc7a3998b",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.layers import Dense, Dropout, Input, Embedding, \\\n",
    "    Concatenate, Reshape, concatenate, Flatten\n",
    "import tensorflow.keras.metrics as km"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "9f1e4f2e-0e45-4e28-8e9c-212e0bb51199",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sba_nn.sba_nn import sbnn_metrics, sbnn_model "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b6cc1a60-d316-4c1b-8f61-c4bd0ec67363",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Imputer object for easy dataset conversion to NN friendly format\n",
    "from sba_nn.sba_nn import sbnn_imputer \n",
    "from sba_nn.sba_nn.sbnn_imputer import NNImputer"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3cbc00d2-bc83-42d9-83eb-65fcf306dccd",
   "metadata": {},
   "source": [
    "## Input Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "6f0ce28e-a80c-415b-b068-5c4d49370da8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'NAICS_orig': 1170, 'rgrp_345': 337, 'rgrp_106': 107, 'rgrp_020': 21}"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "with open(Path(setup.temp_path).joinpath('41_DATA_naics_max_encodings.pkl'), 'rb') as fin:\n",
    "    naics_max_levels= pickle.load(fin)\n",
    "naics_max_levels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "958fb977-c8de-4e6b-8f1b-c8e54acaddb0",
   "metadata": {},
   "outputs": [],
   "source": [
    "grp_features = list(naics_max_levels.keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "e014cd31-0f8d-42e3-82f9-b529eec8a3cc",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "sba_loans = pd.read_parquet(Path(setup.temp_path).joinpath('01_DATA_transformed_nomiss.parquet'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "fca4391e-98cd-44c4-b6fd-313f57e431cc",
   "metadata": {},
   "outputs": [],
   "source": [
    "naics_new = pd.read_parquet(Path(setup.temp_path).joinpath('41_DATA_naics_groups_loans.parquet'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "7066d473-6c62-4407-8862-6e544e61077b",
   "metadata": {},
   "outputs": [],
   "source": [
    "sba_loans = sba_loans.drop(columns=grp_features, errors='ignore') \\\n",
    "    .merge(naics_new[['LoanNr_ChkDgt'] + grp_features], on='LoanNr_ChkDgt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "8e65cb7c-792c-46ff-a530-0d5c64240d60",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(Path(setup.temp_path).joinpath('01_DATA_features.pkl'), 'rb') as fin:\n",
    "    imputer_features = pickle.load(fin)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f69dc43f-e365-4c9e-9ba5-7f163ee26a77",
   "metadata": {},
   "source": [
    "## Modify Data\n",
    "Inject \"1\" values into encodings for the random NAICS hierarchy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "84e89190-1468-459e-8182-d2a47786cf19",
   "metadata": {},
   "outputs": [],
   "source": [
    "for i, n in enumerate(grp_features):\n",
    "    adj_index, _ = train_test_split(sba_loans[sba_loans['dset'] == 'train'].index, \n",
    "                                    train_size=0.10, random_state=645 + i)\n",
    "    sba_loans.loc[adj_index, n] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "8400360c-80bc-4924-b66b-93b6a0050ac0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>NAICS_orig</th>\n",
       "      <th>False</th>\n",
       "      <th>True</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dset</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>test</th>\n",
       "      <td>95552</td>\n",
       "      <td>57496</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train</th>\n",
       "      <td>402164</td>\n",
       "      <td>44684</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>val</th>\n",
       "      <td>95593</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "NAICS_orig   False  True \n",
       "dset                     \n",
       "test         95552  57496\n",
       "train       402164  44684\n",
       "val          95593     11"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Show unseen counts - expect some in train now!\n",
    "pd.crosstab(sba_loans['dset'], sba_loans[grp_features[0]] == 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4d4116da-d689-478a-9c5d-a330f231e615",
   "metadata": {},
   "source": [
    "# Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "f98c6e41-38d0-4d5f-b821-e3837b8c2068",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['NoEmp', 'CreateJob', 'LowDoc', 'DisbursementGross', 'new_business', 'urban_flag', 'franchise_flag', 'missingindicator_LowDoc', 'missingindicator_new_business', 'missingindicator_urban_flag']\n"
     ]
    }
   ],
   "source": [
    "# Numeric features\n",
    "features_numeric = [f for f in imputer_features if 'NAICS' not in f]\n",
    "features_numeric_len = len(features_numeric)\n",
    "print(features_numeric)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "a74fa386-4011-45ee-bdb5-d671db9ffa91",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Use analagous embedding dimensions\n",
    "features_naics_hier = ['NAICS_4', 'NAICS_3', 'NAICS_sector']\n",
    "features_grp_all = ['NAICS'] + features_naics_hier\n",
    "grp_emb_dim = [setup.nn_naics_embed_size_dict[n] for n in features_grp_all]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "867031ee-8b6e-424d-b497-8e19409ba1d9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'NAICS_orig': 1170, 'rgrp_345': 337, 'rgrp_106': 107, 'rgrp_020': 21}\n",
      "[8, 4, 4, 2]\n"
     ]
    }
   ],
   "source": [
    "print(naics_max_levels)\n",
    "print(grp_emb_dim)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b979396c-c822-4940-a292-6e07f1b54f58",
   "metadata": {},
   "source": [
    "##### Datasets for train, validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "43bca011-a59d-4496-9f03-87f4e350a253",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = sba_loans[['dset', 'LoanNr_ChkDgt'] + features_numeric].set_index('LoanNr_ChkDgt').sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "aa162d01-ed73-4ae3-832e-4aeea09b70f4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "training X: (446848, 10), y:(446848,)\n"
     ]
    }
   ],
   "source": [
    "X_train = X[X['dset'] == 'train'].drop(columns='dset')\n",
    "y_train = sba_loans[sba_loans['dset'] == 'train'].set_index('LoanNr_ChkDgt').sort_index()['target']\n",
    "print(f'training X: {X_train.shape}, y:{y_train.shape}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "b32b5fd3-6f6e-4aec-8e20-0e2b71a4814c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val X: (95604, 10), y:(95604,)\n"
     ]
    }
   ],
   "source": [
    "X_val = X[X['dset'] == 'val'].drop(columns='dset')\n",
    "y_val = sba_loans[sba_loans['dset'] == 'val'].set_index('LoanNr_ChkDgt').sort_index()['target']\n",
    "print(f'val X: {X_val.shape}, y:{y_val.shape}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "01b8ff96-159b-4283-b5fb-c4b5dbd277d5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.20230592953308507\n"
     ]
    }
   ],
   "source": [
    "base_thresh = y_train.mean()\n",
    "print(base_thresh)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "1b8ce4c9-6907-4b70-982a-b92a52b3159d",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_hier = [sba_loans[['dset', 'LoanNr_ChkDgt', n]].set_index('LoanNr_ChkDgt').sort_index() \\\n",
    "           for n in grp_features]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "a6f129e0-a4d0-4d93-9886-153c9803d403",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_hier_train = [Xn[Xn['dset'] == 'train'].drop(columns='dset') for Xn in X_hier]\n",
    "X_hier_val =  [Xn[Xn['dset'] == 'val'].drop(columns='dset') for Xn in X_hier]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3db4ea1c-77b2-4721-b106-1e66fc502139",
   "metadata": {},
   "source": [
    "## Create, fit model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "4d48ddfa-580c-4321-9917-0bd176b67d87",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024-05-29 17:30:34.069952: I metal_plugin/src/device/metal_device.cc:1154] Metal device set to: Apple M1\n",
      "2024-05-29 17:30:34.069993: I metal_plugin/src/device/metal_device.cc:296] systemMemory: 16.00 GB\n",
      "2024-05-29 17:30:34.070017: I metal_plugin/src/device/metal_device.cc:313] maxCacheSize: 5.33 GB\n",
      "2024-05-29 17:30:34.070101: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:303] Could not identify NUMA node of platform GPU ID 0, defaulting to 0. Your kernel may not have been built with NUMA support.\n",
      "2024-05-29 17:30:34.070262: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:269] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 0 MB memory) -> physical PluggableDevice (device: 0, name: METAL, pci bus id: <undefined>)\n"
     ]
    }
   ],
   "source": [
    "this_model = sbnn_model.create_emb_model(features_numeric_len,\n",
    "                                         list(naics_max_levels.values()),\n",
    "                                         grp_emb_dim,\n",
    "                                         naics_embedding_names = grp_features,\n",
    "                                         hidden_size = setup.nn_layer_sizes,\n",
    "                                         activation='tanh',\n",
    "                                         lr=setup.nn_learning_rate,\n",
    "                                         opt_func = setup.nn_optimizer,\n",
    "                                         dropout = setup.nn_dropout)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "47036621-ef13-4957-997b-f8ef86016f7f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "__________________________________________________________________________________________________\n",
      " Layer (type)                Output Shape                 Param #   Connected to                  \n",
      "==================================================================================================\n",
      " input_naics_00 (InputLayer  [(None, 1)]                  0         []                            \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " input_naics_01 (InputLayer  [(None, 1)]                  0         []                            \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " input_naics_02 (InputLayer  [(None, 1)]                  0         []                            \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " input_naics_03 (InputLayer  [(None, 1)]                  0         []                            \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " emb_naics_00 (Embedding)    (None, 1, 8)                 9360      ['input_naics_00[0][0]']      \n",
      "                                                                                                  \n",
      " emb_naics_01 (Embedding)    (None, 1, 4)                 1348      ['input_naics_01[0][0]']      \n",
      "                                                                                                  \n",
      " emb_naics_02 (Embedding)    (None, 1, 4)                 428       ['input_naics_02[0][0]']      \n",
      "                                                                                                  \n",
      " emb_naics_03 (Embedding)    (None, 1, 2)                 42        ['input_naics_03[0][0]']      \n",
      "                                                                                                  \n",
      " NAICS_orig (Flatten)        (None, 8)                    0         ['emb_naics_00[0][0]']        \n",
      "                                                                                                  \n",
      " rgrp_345 (Flatten)          (None, 4)                    0         ['emb_naics_01[0][0]']        \n",
      "                                                                                                  \n",
      " rgrp_106 (Flatten)          (None, 4)                    0         ['emb_naics_02[0][0]']        \n",
      "                                                                                                  \n",
      " rgrp_020 (Flatten)          (None, 2)                    0         ['emb_naics_03[0][0]']        \n",
      "                                                                                                  \n",
      " input_numeric (InputLayer)  [(None, 10)]                 0         []                            \n",
      "                                                                                                  \n",
      " input_concat (Concatenate)  (None, 28)                   0         ['NAICS_orig[0][0]',          \n",
      "                                                                     'rgrp_345[0][0]',            \n",
      "                                                                     'rgrp_106[0][0]',            \n",
      "                                                                     'rgrp_020[0][0]',            \n",
      "                                                                     'input_numeric[0][0]']       \n",
      "                                                                                                  \n",
      " layer_00 (Dense)            (None, 128)                  3712      ['input_concat[0][0]']        \n",
      "                                                                                                  \n",
      " dropout_00 (Dropout)        (None, 128)                  0         ['layer_00[0][0]']            \n",
      "                                                                                                  \n",
      " layer_01 (Dense)            (None, 64)                   8256      ['dropout_00[0][0]']          \n",
      "                                                                                                  \n",
      " dropout_01 (Dropout)        (None, 64)                   0         ['layer_01[0][0]']            \n",
      "                                                                                                  \n",
      " output (Dense)              (None, 1)                    65        ['dropout_01[0][0]']          \n",
      "                                                                                                  \n",
      "==================================================================================================\n",
      "Total params: 23211 (90.67 KB)\n",
      "Trainable params: 23211 (90.67 KB)\n",
      "Non-trainable params: 0 (0.00 Byte)\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "this_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "cbb85f7c-127b-4a78-bcd0-bc00cc6bed14",
   "metadata": {},
   "outputs": [],
   "source": [
    "callback = keras.callbacks.EarlyStopping(patience=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "31772016-3643-468b-8d65-c456168cad6e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024-05-29 17:30:35.956313: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:114] Plugin optimizer for device_type GPU is enabled.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13961/13964 [============================>.] - ETA: 0s - loss: 0.4630 - auc: 0.3344 - auc_1: 0.6981"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024-05-29 17:34:41.791814: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:114] Plugin optimizer for device_type GPU is enabled.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13964/13964 [==============================] - 269s 19ms/step - loss: 0.4630 - auc: 0.3344 - auc_1: 0.6981 - val_loss: 0.4554 - val_auc: 0.3718 - val_auc_1: 0.7238\n",
      "Epoch 2/20\n",
      "13964/13964 [==============================] - 265s 19ms/step - loss: 0.4549 - auc: 0.3572 - auc_1: 0.7158 - val_loss: 0.4520 - val_auc: 0.3771 - val_auc_1: 0.7268\n",
      "Epoch 3/20\n",
      "13964/13964 [==============================] - 272s 20ms/step - loss: 0.4528 - auc: 0.3642 - auc_1: 0.7202 - val_loss: 0.4515 - val_auc: 0.3799 - val_auc_1: 0.7287\n",
      "Epoch 4/20\n",
      "13964/13964 [==============================] - 279s 20ms/step - loss: 0.4517 - auc: 0.3684 - auc_1: 0.7226 - val_loss: 0.4503 - val_auc: 0.3820 - val_auc_1: 0.7293\n",
      "Epoch 5/20\n",
      "13964/13964 [==============================] - 270s 19ms/step - loss: 0.4507 - auc: 0.3715 - auc_1: 0.7244 - val_loss: 0.4507 - val_auc: 0.3807 - val_auc_1: 0.7289\n",
      "Epoch 6/20\n",
      "13964/13964 [==============================] - 258s 18ms/step - loss: 0.4499 - auc: 0.3749 - auc_1: 0.7260 - val_loss: 0.4508 - val_auc: 0.3815 - val_auc_1: 0.7298\n",
      "Epoch 7/20\n",
      "13964/13964 [==============================] - 254s 18ms/step - loss: 0.4495 - auc: 0.3758 - auc_1: 0.7270 - val_loss: 0.4504 - val_auc: 0.3842 - val_auc_1: 0.7302\n"
     ]
    }
   ],
   "source": [
    "this_history = this_model.fit([X_train] + X_hier_train, y_train,\n",
    "                              validation_data=([X_val] + X_hier_val, y_val),\n",
    "                              batch_size=setup.nn_batch_size,\n",
    "                              epochs=setup.nn_epochs,\n",
    "                              callbacks=[callback])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "460c4b43-c673-4ee1-b070-d2d596ac4561",
   "metadata": {},
   "outputs": [],
   "source": [
    "this_history_df = sbnn_model.process_history(this_history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "ad4783f2-8d0b-4850-a45a-3a4be8e63318",
   "metadata": {},
   "outputs": [],
   "source": [
    "this_history_df.to_csv(Path(setup.temp_path).joinpath('42_REPORT_fit_history.csv'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "73e89efd-75a9-4b27-81a8-13e12d5482be",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "this_history_df[['loss', 'val_loss']].plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "2342a31f-4891-4834-851a-b90b8b729c3c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "this_history_df[['auc', 'val_auc']].plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "cee80637-cabf-4974-b476-08c5fba165ff",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "this_history_df[['auc_roc', 'val_auc_roc']].plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "e184b0ca-1d31-4798-88a0-4e590b67062e",
   "metadata": {},
   "outputs": [],
   "source": [
    "this_model.save(Path(setup.temp_path).joinpath('42_DATA_model.keras'),save_format='tf')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d326c9ea-4759-4418-a64d-56c151049c25",
   "metadata": {},
   "source": [
    "## Predictions on all data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "d5c39a5a-dc19-4606-b744-34865bc6c5c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_hier = [sba_loans[['dset', 'LoanNr_ChkDgt', n]].set_index('LoanNr_ChkDgt').sort_index() \\\n",
    "           for n in grp_features]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "cd6910ee-5ffe-4def-9aa8-7dba76446b08",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   40/21735 [..............................] - ETA: 58s "
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024-05-29 18:01:42.169966: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:114] Plugin optimizer for device_type GPU is enabled.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "21735/21735 [==============================] - 57s 3ms/step\n"
     ]
    }
   ],
   "source": [
    "all_predictions = this_model.predict([X.drop(columns='dset')] +\n",
    "                                     [Xn.drop(columns='dset') for Xn in X_hier])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "98a2cce1-b7d3-4ebc-8869-80705297e145",
   "metadata": {},
   "outputs": [],
   "source": [
    "all_predictions_df = pd.DataFrame(all_predictions, index=X.index) \\\n",
    "    .set_axis(['predict_prob'], axis=1) \\\n",
    "    .reset_index() \\\n",
    "    .merge(sba_loans[['target', 'LoanNr_ChkDgt', 'dset', 'dset_naics_holdout', 'NAICS']], \n",
    "           on='LoanNr_ChkDgt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "3fc7da9c-6bd2-4d1b-b77f-b62679825d5d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>predict_prob</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>predict_prob</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.319962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>target</th>\n",
       "      <td>0.319962</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              predict_prob    target\n",
       "predict_prob      1.000000  0.319962\n",
       "target            0.319962  1.000000"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_predictions_df[['predict_prob', 'target']].corr(method='spearman')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e79afc4e-cb67-4064-9b19-dc52500baef2",
   "metadata": {},
   "source": [
    "##### Threshold Tune & Binary Predictions\n",
    "Using training probability predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "ccd0bc23-fbc0-40d0-8d8b-f2302405a875",
   "metadata": {},
   "outputs": [],
   "source": [
    "all_pred_train = all_predictions_df[all_predictions_df['dset'] == 'train']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "df36d9e9-f35e-46b3-a5be-08081f998319",
   "metadata": {},
   "outputs": [],
   "source": [
    "thresh_tune_data = sbnn_metrics.get_f1_frame(all_pred_train['target'], \n",
    "                                        all_pred_train['predict_prob'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "de634156-caf6-4a52-b5d7-44b8e986aa1d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>thresh</th>\n",
       "      <th>f1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.26</td>\n",
       "      <td>0.451808</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.24</td>\n",
       "      <td>0.451675</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.22</td>\n",
       "      <td>0.450018</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    thresh        f1\n",
       "13    0.26  0.451808\n",
       "12    0.24  0.451675\n",
       "11    0.22  0.450018"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "thresh_tune_data.sort_values('f1', ascending=False, inplace=True)\n",
    "thresh_tune_data.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "c746b880-d781-4bc1-9df2-89211127c22a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.26"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best_thresh = thresh_tune_data['thresh'].iloc[0]\n",
    "best_thresh"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3c0ac5e2-be0d-42f0-960d-1f81b4a5d4e2",
   "metadata": {},
   "source": [
    "##### Append binary predictions to probability predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "641da72e-1bca-4fe7-96c4-6c029b521bd7",
   "metadata": {},
   "outputs": [],
   "source": [
    "all_predictions_df['predict_bin'] = sbnn_metrics.get_binary_predictions(all_predictions_df['predict_prob'], best_thresh)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "9f6ecdf4-ac01-4c39-8315-e8583daa337e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "predict_bin\n",
       "0    0.612394\n",
       "1    0.387606\n",
       "Name: proportion, dtype: float64"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_predictions_df['predict_bin'].value_counts(normalize=True, dropna=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "cbfed830-7ff8-40f6-a691-6b894b26d6bc",
   "metadata": {},
   "outputs": [],
   "source": [
    "all_predictions_df.to_parquet(Path(setup.temp_path).joinpath('42_DATA_predictions.parquet'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eb0ecbd2-abff-46d8-8c70-f7bccd363b69",
   "metadata": {},
   "source": [
    "## Metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "be4ceefe-c7a8-45bd-8573-2daefe99a1c2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>dset</th>\n",
       "      <th>accuracy_score</th>\n",
       "      <th>f1_score</th>\n",
       "      <th>precision_score</th>\n",
       "      <th>recall_score</th>\n",
       "      <th>average_precision_score</th>\n",
       "      <th>roc_auc_score</th>\n",
       "      <th>dset_naics_holdout</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>test</td>\n",
       "      <td>0.676735</td>\n",
       "      <td>0.436638</td>\n",
       "      <td>0.336670</td>\n",
       "      <td>0.621048</td>\n",
       "      <td>0.365615</td>\n",
       "      <td>0.719027</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>train</td>\n",
       "      <td>0.674245</td>\n",
       "      <td>0.451808</td>\n",
       "      <td>0.342511</td>\n",
       "      <td>0.663551</td>\n",
       "      <td>0.388053</td>\n",
       "      <td>0.733548</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>val</td>\n",
       "      <td>0.672869</td>\n",
       "      <td>0.451171</td>\n",
       "      <td>0.342818</td>\n",
       "      <td>0.659671</td>\n",
       "      <td>0.384322</td>\n",
       "      <td>0.730243</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>test</td>\n",
       "      <td>0.670950</td>\n",
       "      <td>0.446514</td>\n",
       "      <td>0.339172</td>\n",
       "      <td>0.653258</td>\n",
       "      <td>0.378486</td>\n",
       "      <td>0.725823</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>test</td>\n",
       "      <td>0.686351</td>\n",
       "      <td>0.418548</td>\n",
       "      <td>0.331884</td>\n",
       "      <td>0.566466</td>\n",
       "      <td>0.342861</td>\n",
       "      <td>0.708391</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    dset  accuracy_score  f1_score  precision_score  recall_score  \\\n",
       "0   test        0.676735  0.436638         0.336670      0.621048   \n",
       "1  train        0.674245  0.451808         0.342511      0.663551   \n",
       "2    val        0.672869  0.451171         0.342818      0.659671   \n",
       "0   test        0.670950  0.446514         0.339172      0.653258   \n",
       "1   test        0.686351  0.418548         0.331884      0.566466   \n",
       "\n",
       "   average_precision_score  roc_auc_score  dset_naics_holdout  \n",
       "0                 0.365615       0.719027                 NaN  \n",
       "1                 0.388053       0.733548                 NaN  \n",
       "2                 0.384322       0.730243                 NaN  \n",
       "0                 0.378486       0.725823                 0.0  \n",
       "1                 0.342861       0.708391                 1.0  "
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics_dset_df = all_predictions_df.groupby('dset') \\\n",
    "    .apply(lambda x: sbnn_metrics.dset_metrics(x.target, x.predict_bin, x.predict_prob)) \\\n",
    "    .reset_index()\n",
    "metrics_test_df = all_predictions_df[all_predictions_df['dset'] == 'test'] \\\n",
    "    .groupby(['dset', 'dset_naics_holdout']) \\\n",
    "    .apply(lambda x: sbnn_metrics.dset_metrics(x.target, x.predict_bin, x.predict_prob)) \\\n",
    "    .reset_index()\n",
    "metrics_df = pd.concat([metrics_dset_df, metrics_test_df])\n",
    "metrics_df.to_csv(Path(setup.temp_path).joinpath('42_REPORT_metrics.csv'), index=True)\n",
    "metrics_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "52e382b5-e9dc-47a1-9bf4-6d0b6b4a0c61",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6f69cf93-02e2-413b-8f2b-fcdb908057d3",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "tf_p39",
   "language": "python",
   "name": "tf_p39"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}