315 lines
204 KiB
Text
315 lines
204 KiB
Text
|
{
|
||
|
|
"cells": [
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"id": "717f94d3-edfa-4122-9902-212a3456bb8c",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"[EDA](eda.ipynb)"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 1,
|
||
|
|
"id": "c47122c5-bdcd-4b6b-8d22-8958ad910eca",
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-21T20:30:04.392164Z",
|
||
|
|
"iopub.status.busy": "2022-07-21T20:30:04.391759Z",
|
||
|
|
"iopub.status.idle": "2022-07-21T20:30:06.057099Z",
|
||
|
|
"shell.execute_reply": "2022-07-21T20:30:06.056318Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-21T20:30:04.392085Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [],
|
||
|
|
"source": [
|
||
|
|
"import numpy as np\n",
|
||
|
|
"import pandas as pd\n",
|
||
|
|
"import matplotlib.pyplot as plt\n",
|
||
|
|
"from sklearn.model_selection import train_test_split\n",
|
||
|
|
"from sklearn.pipeline import Pipeline\n",
|
||
|
|
"from sklearn.preprocessing import (StandardScaler,\n",
|
||
|
|
" QuantileTransformer,\n",
|
||
|
|
" Normalizer,\n",
|
||
|
|
" MinMaxScaler,\n",
|
||
|
|
" RobustScaler,\n",
|
||
|
|
" PowerTransformer)\n",
|
||
|
|
"\n",
|
||
|
|
"from sklearn.linear_model import (Lars,\n",
|
||
|
|
" Ridge,\n",
|
||
|
|
" Lasso,\n",
|
||
|
|
" LarsCV,\n",
|
||
|
|
" LassoCV,\n",
|
||
|
|
" RidgeCV,\n",
|
||
|
|
" LassoLars,\n",
|
||
|
|
" ElasticNet,\n",
|
||
|
|
" LassoLarsCV,\n",
|
||
|
|
" LassoLarsIC,\n",
|
||
|
|
" ElasticNetCV,\n",
|
||
|
|
" SGDRegressor,\n",
|
||
|
|
" LinearRegression,\n",
|
||
|
|
" OrthogonalMatchingPursuit,\n",
|
||
|
|
" OrthogonalMatchingPursuitCV)\n",
|
||
|
|
"\n",
|
||
|
|
"from sklearn.neighbors import KNeighborsRegressor\n",
|
||
|
|
"from sklearn.svm import LinearSVR\n",
|
||
|
|
"from sklearn.metrics import mean_squared_error, r2_score"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 2,
|
||
|
|
"id": "8c3b98e1-9a8f-4d42-8a29-5d030934a4ed",
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-21T20:30:06.059480Z",
|
||
|
|
"iopub.status.busy": "2022-07-21T20:30:06.059058Z",
|
||
|
|
"iopub.status.idle": "2022-07-21T20:30:11.086526Z",
|
||
|
|
"shell.execute_reply": "2022-07-21T20:30:11.085657Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-21T20:30:06.059453Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"image/png": "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
|
||
|
|
"text/plain": [
|
||
|
|
"<Figure size 720x432 with 1 Axes>"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"metadata": {
|
||
|
|
"needs_background": "light"
|
||
|
|
},
|
||
|
|
"output_type": "display_data"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"name": "stdout",
|
||
|
|
"output_type": "stream",
|
||
|
|
"text": [
|
||
|
|
"| R2 Test | R2 Train | RMSE Test | RMSE Train |\n",
|
||
|
|
"|-----------|----------|-------------|------------|\n",
|
||
|
|
"| Min: 0.62 | Min:0.72 | Min: 3.24 | Min:3.57 |\n",
|
||
|
|
"| Avg: 0.73 | Avg:0.75 | Avg: 4.01 | Avg:3.93 |\n",
|
||
|
|
"| Max: 0.82 | Max:0.78 | Max: 5.04 | Max:4.15 |\n",
|
||
|
|
"\n"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"image/png": "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
|
||
|
|
"text/plain": [
|
||
|
|
"<Figure size 720x360 with 1 Axes>"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"metadata": {
|
||
|
|
"needs_background": "light"
|
||
|
|
},
|
||
|
|
"output_type": "display_data"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"name": "stdout",
|
||
|
|
"output_type": "stream",
|
||
|
|
"text": [
|
||
|
|
"Truck Avg: 16.23\n",
|
||
|
|
"Mustang Avg: 20.68\n",
|
||
|
|
"Grom Avg: 36.07\n",
|
||
|
|
"Burb Avg: 15.86\n",
|
||
|
|
"Corolla Avg: 21.06\n",
|
||
|
|
"Steve Avg: 16.37\n"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"X = pd.read_csv('data/X.csv')\n",
|
||
|
|
"y = pd.read_csv('data/y.csv').mpg\n",
|
||
|
|
"\n",
|
||
|
|
"eff = 1\n",
|
||
|
|
"\n",
|
||
|
|
"v6_s197_05_hp = 216\n",
|
||
|
|
"v6_s197_05_ci = 245\n",
|
||
|
|
"v6_s197_05_cl = 6\n",
|
||
|
|
"v6_s197_05_weight = 3300\n",
|
||
|
|
"v6_s197_05_eff = (v6_s197_05_hp/\\\n",
|
||
|
|
" v6_s197_05_ci)*eff\n",
|
||
|
|
"v6_s197_05 = {'horsepower':v6_s197_05_hp,\n",
|
||
|
|
" 'bore_size':v6_s197_05_ci/v6_s197_05_cl,\n",
|
||
|
|
" 'grunt':(v6_s197_05_ci/v6_s197_05_cl)/v6_s197_05_eff,\n",
|
||
|
|
" 'load':v6_s197_05_ci/v6_s197_05_weight}\n",
|
||
|
|
"mustang_hp = 400\n",
|
||
|
|
"mustang_ci = 302\n",
|
||
|
|
"mustang_cl = 8\n",
|
||
|
|
"mustang_weight = 3600\n",
|
||
|
|
"mustang_eff = (mustang_hp/\\\n",
|
||
|
|
" mustang_ci)*eff\n",
|
||
|
|
"mustang = {'horsepower':mustang_hp,\n",
|
||
|
|
" 'bore_size':mustang_ci/mustang_cl,\n",
|
||
|
|
" 'grunt':(mustang_ci/mustang_cl)/mustang_eff,\n",
|
||
|
|
" 'load':mustang_ci/mustang_weight}\n",
|
||
|
|
"corolla_hp = 140\n",
|
||
|
|
"corolla_ci = 110\n",
|
||
|
|
"corolla_cl = 4\n",
|
||
|
|
"corolla_weight = 2800\n",
|
||
|
|
"corolla_eff = (corolla_hp/\\\n",
|
||
|
|
" corolla_ci)*eff\n",
|
||
|
|
"corolla = {'horsepower':corolla_hp,\n",
|
||
|
|
" 'bore_size':corolla_ci/corolla_cl,\n",
|
||
|
|
" 'grunt':(corolla_ci/corolla_cl)/corolla_eff,\n",
|
||
|
|
" 'load':corolla_ci/corolla_weight}\n",
|
||
|
|
"truck_hp = 500\n",
|
||
|
|
"truck_ci = 359\n",
|
||
|
|
"truck_cl = 6\n",
|
||
|
|
"truck_weight = 6500\n",
|
||
|
|
"truck_eff = (truck_hp/\\\n",
|
||
|
|
" truck_ci)*eff\n",
|
||
|
|
"truck = {'horsepower':truck_hp,\n",
|
||
|
|
" 'bore_size':truck_ci/truck_cl,\n",
|
||
|
|
" 'grunt':(truck_ci/truck_cl)/truck_eff,\n",
|
||
|
|
" 'load':truck_ci/truck_weight}\n",
|
||
|
|
"grom_hp = 12\n",
|
||
|
|
"grom_ci = 7.6\n",
|
||
|
|
"grom_cl = 1\n",
|
||
|
|
"grom_weight = 400\n",
|
||
|
|
"grom_eff = (grom_hp/\\\n",
|
||
|
|
" grom_ci)*eff\n",
|
||
|
|
"grom = {'horsepower':grom_hp,\n",
|
||
|
|
" 'bore_size':grom_ci/grom_cl,\n",
|
||
|
|
" 'grunt':(grom_ci/grom_cl)/grom_eff,\n",
|
||
|
|
" 'load':grom_ci/grom_weight}\n",
|
||
|
|
"burb_hp = 320\n",
|
||
|
|
"burb_ci = 325\n",
|
||
|
|
"burb_cl = 8\n",
|
||
|
|
"burb_weight = 6000\n",
|
||
|
|
"burb_eff = (burb_hp/\\\n",
|
||
|
|
" burb_ci)*eff\n",
|
||
|
|
"burb = {'horsepower':burb_hp,\n",
|
||
|
|
" 'bore_size':burb_ci/burb_cl,\n",
|
||
|
|
" 'grunt':(burb_ci/burb_cl)/burb_eff,\n",
|
||
|
|
" 'load':burb_ci/burb_weight}\n",
|
||
|
|
"\n",
|
||
|
|
"mdf = pd.DataFrame(mustang,index=[0])\n",
|
||
|
|
"cdf = pd.DataFrame(corolla,index=[0])\n",
|
||
|
|
"tdf = pd.DataFrame(truck,index=[0])\n",
|
||
|
|
"gdf = pd.DataFrame(grom,index=[0])\n",
|
||
|
|
"bdf = pd.DataFrame(burb,index=[0])\n",
|
||
|
|
"sm5 = pd.DataFrame(v6_s197_05,index=[0])\n",
|
||
|
|
"\n",
|
||
|
|
"mustang_predicts = []\n",
|
||
|
|
"corolla_predicts = []\n",
|
||
|
|
"truck_predicts = []\n",
|
||
|
|
"grom_predicts = []\n",
|
||
|
|
"burb_predicts = []\n",
|
||
|
|
"v6_s197_05_predicts = []\n",
|
||
|
|
"\n",
|
||
|
|
"r2_test_list = []\n",
|
||
|
|
"r2_train_list = []\n",
|
||
|
|
"rmse_test_list = []\n",
|
||
|
|
"rmse_train_list = []\n",
|
||
|
|
"\n",
|
||
|
|
"for i in range(201):\n",
|
||
|
|
" X_train, X_test, y_train, y_test = train_test_split(X, y)\n",
|
||
|
|
"\n",
|
||
|
|
" pipe = Pipeline([\n",
|
||
|
|
" # ('minmax', MinMaxScaler()),\n",
|
||
|
|
" # ('ss', StandardScaler()),\n",
|
||
|
|
" ('qt', QuantileTransformer(n_quantiles=297)),\n",
|
||
|
|
" # ('rob', RobustScaler()),\n",
|
||
|
|
" \n",
|
||
|
|
" ('linreg', LinearRegression()),\n",
|
||
|
|
" # ('lasso', Lasso()),\n",
|
||
|
|
" # ('lassocv', LassoCV()),\n",
|
||
|
|
" # ('ridge', Ridge()),\n",
|
||
|
|
" # ('ridgeCV', RidgeCV),\n",
|
||
|
|
" # ('lsvr', LinearSVR())\n",
|
||
|
|
" ])\n",
|
||
|
|
"\n",
|
||
|
|
" model = pipe.fit(X_train,y_train)\n",
|
||
|
|
" test_predict = model.predict(X_test)\n",
|
||
|
|
" train_predict = model.predict(X_train)\n",
|
||
|
|
"\n",
|
||
|
|
" r2_test = r2_score(y_test, test_predict)\n",
|
||
|
|
" r2_train = r2_score(y_train, train_predict)\n",
|
||
|
|
" rmse_test = mean_squared_error(y_test, test_predict ,squared=False)\n",
|
||
|
|
" rmse_train = mean_squared_error(y_train, train_predict ,squared=False)\n",
|
||
|
|
"\n",
|
||
|
|
" r2_test_list.append(r2_test)\n",
|
||
|
|
" r2_train_list.append(r2_train)\n",
|
||
|
|
" rmse_test_list.append(rmse_test)\n",
|
||
|
|
" rmse_train_list.append(rmse_train)\n",
|
||
|
|
" truck_predicts.append(model.predict(tdf)[0])\n",
|
||
|
|
" mustang_predicts.append(model.predict(mdf)[0])\n",
|
||
|
|
" grom_predicts.append(model.predict(gdf)[0])\n",
|
||
|
|
" burb_predicts.append(model.predict(bdf)[0])\n",
|
||
|
|
" corolla_predicts.append(model.predict(cdf)[0])\n",
|
||
|
|
" v6_s197_05_predicts.append(model.predict(sm5)[0])\n",
|
||
|
|
"\n",
|
||
|
|
"plt.subplots(figsize=(10,6))\n",
|
||
|
|
"plt.title('R-squared over 200 iterations')\n",
|
||
|
|
"plt.plot(r2_test_list,label='R2 Test')\n",
|
||
|
|
"plt.plot(r2_train_list,label='R2 Train')\n",
|
||
|
|
"plt.legend()\n",
|
||
|
|
"plt.show();\n",
|
||
|
|
"\n",
|
||
|
|
"avg = np.mean\n",
|
||
|
|
"print(f'''| R2 Test | R2 Train | RMSE Test | RMSE Train |\n",
|
||
|
|
"|-----------|----------|-------------|------------|\n",
|
||
|
|
"| Min: {min(r2_test_list):.2f} | Min:{min(r2_train_list):.2f} | Min: {min(rmse_test_list):.2f} | Min:{min(rmse_train_list):.2f} |\n",
|
||
|
|
"| Avg: {avg(r2_test_list):.2f} | Avg:{avg(r2_train_list):.2f} | Avg: {avg(rmse_test_list):.2f} | Avg:{avg(rmse_train_list):.2f} |\n",
|
||
|
|
"| Max: {max(r2_test_list):.2f} | Max:{max(r2_train_list):.2f} | Max: {max(rmse_test_list):.2f} | Max:{max(rmse_train_list):.2f} |\n",
|
||
|
|
"''')\n",
|
||
|
|
"plt.subplots(figsize=(10,5))\n",
|
||
|
|
"plt.title('RMSE over 200 iterations')\n",
|
||
|
|
"plt.plot(rmse_test_list,label='RMSE Test')\n",
|
||
|
|
"plt.plot(rmse_train_list,label='RMSE Train')\n",
|
||
|
|
"plt.legend()\n",
|
||
|
|
"plt.show();\n",
|
||
|
|
"\n",
|
||
|
|
"print(f'Truck Avg: {avg(truck_predicts):.2f}')\n",
|
||
|
|
"print(f'Mustang Avg: {avg(mustang_predicts):.2f}')\n",
|
||
|
|
"print(f'Grom Avg: {avg(grom_predicts):.2f}')\n",
|
||
|
|
"print(f'Burb Avg: {avg(burb_predicts):.2f}')\n",
|
||
|
|
"print(f'Corolla Avg: {avg(corolla_predicts):.2f}')\n",
|
||
|
|
"print(f'Steve Avg: {avg(v6_s197_05_predicts):.2f}')"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"id": "67f5823f-4d25-4e75-95a0-b11249d861e8",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"After testing lots of models, scalers, and experimenting with different mixtures of features, I've come to the conclusion that it's futile.\n",
|
||
|
|
"\n",
|
||
|
|
"It's cliché, but at under 400 rows I think I can complain about not having enough data.\n",
|
||
|
|
"\n",
|
||
|
|
"QuantileTransformer is the heavy lifter and the model really doesn't seem to matter so long as it's linear. In fact, the scores go down as the models get fancier. I think because it's overthinking such a small data set, it's making connections that are coincidental.\n",
|
||
|
|
"\n",
|
||
|
|
"The best metric I found was to make predictions on my own vehicles, in other words it's unseen data that I could just pull from nothing. I had used model year as a feature and it seemed to perform quite well until I introduced a 2012, it predicted MPG into the hundreds. I think model year would be a great feature if the data spanned across decades and there was good representation. It could probably even be made categorical as it would be an indicator of tech advancing over decades. All of the vehicles here are more or less in the same \"tech era\" but the model did seem to find a signal. At any rate it's too unstable if used with data outside the training set.\n",
|
||
|
|
"\n",
|
||
|
|
"My personal vehicle predictions are actually quite close. They're a bit on the lower end, but also consider that neither one is even close to being represented in the training data. My truck is a turbo diesel (boost to efficiency) and my car is gas but has some fancy cam phasing and electronic fuel injection. Basically they have higher efficiency compared to anything in the training set."
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"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": "3.10.5"
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"nbformat": 4,
|
||
|
|
"nbformat_minor": 5
|
||
|
|
}
|