534 lines
148 KiB
Plaintext
534 lines
148 KiB
Plaintext
{
|
|
"cells": [
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {
|
|
"id": "pbXYV6QX4In0"
|
|
},
|
|
"source": [
|
|
"# Would you survive Titanic?\n",
|
|
"\n",
|
|
"RMS Titanic was a British passenger ship that sank in the North\n",
|
|
"Atlantic Ocean in the early morning hours of April 15, 1912, after\n",
|
|
"striking an iceberg during her maiden voyage. The majority of its\n",
|
|
"passengers died in the accident. As we will discover, not all of them\n",
|
|
"had the same chance to survive.\n",
|
|
"\n",
|
|
"Data about 887 passengers have been collected and randomly divided\n",
|
|
"into a training and a test set. The training set includes 710 samples\n",
|
|
"and is stored in the file `titanic-train.txt`, while the test set\n",
|
|
"is composed of 177 cases and is stored in `titanic-test.txt`.\n",
|
|
"Each row in the files represents a different passenger, and reports the\n",
|
|
"following features:\n",
|
|
"- the ticket class (1st, 2nd or 3rd);\n",
|
|
"- sex ($0 \\to$ male, $1 \\to$ female);\n",
|
|
"- age, in years;\n",
|
|
"- number of siblings and spouses aboard;\n",
|
|
"- number of parents and children aboard;\n",
|
|
"- the passenger fare.\n",
|
|
"\n",
|
|
"The last column reports whether the passenger survived (1) or not (0).\n",
|
|
"The files can be obtained by executing the following cell."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {
|
|
"id": "Mmsh-_Kg7DPY"
|
|
},
|
|
"source": [
|
|
"Load and correctly shape training data:"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"metadata": {
|
|
"colab": {
|
|
"base_uri": "https://localhost:8080/"
|
|
},
|
|
"id": "Bfpc8l8F6OGy",
|
|
"outputId": "84efe5d7-8ad8-4ece-eeab-72d492d86d45"
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stderr",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Matplotlib is building the font cache; this may take a moment.\n"
|
|
]
|
|
},
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"torch.Size([710, 6]) torch.float32\n",
|
|
"torch.Size([710]) torch.int64\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"import torch\n",
|
|
"import matplotlib.pyplot as plt\n",
|
|
"\n",
|
|
"f = open(\"titanic-train.txt\")\n",
|
|
"data = [float(x) for x in f.read().split()]\n",
|
|
"f.close()\n",
|
|
"\n",
|
|
"# we change the shape to have 7 columns and whatever rows (we are restoring the shape in the file)\n",
|
|
"data = torch.tensor(data).view(-1, 7)\n",
|
|
"# The first 6 are features, the last one is the label\n",
|
|
"X = data[:, :6]\n",
|
|
"Y = data[:, 6].long()\n",
|
|
"\n",
|
|
"# features are float, label is int\n",
|
|
"print(X.shape, X.dtype)\n",
|
|
"print(Y.shape, Y.dtype)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {
|
|
"id": "xMKCVhbe7Q9b"
|
|
},
|
|
"source": [
|
|
"## Training a model\n",
|
|
"\n",
|
|
"Inference function definition:"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 3,
|
|
"metadata": {
|
|
"id": "8U5tPs-u7hjP"
|
|
},
|
|
"outputs": [],
|
|
"source": [
|
|
"def logreg_inference(w, b, X):\n",
|
|
" return torch.sigmoid(X @ w + b)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {
|
|
"id": "46K5bBBx7xL0"
|
|
},
|
|
"source": [
|
|
"Training loop:"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 14,
|
|
"metadata": {
|
|
"id": "VfB0Zesn7uVY"
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"0 0.6931472420692444 0.6211267709732056\n",
|
|
"100 0.5695157051086426 0.7070422768592834\n",
|
|
"200 0.5391784906387329 0.7281690239906311\n",
|
|
"300 0.5169594287872314 0.7464788556098938\n",
|
|
"400 0.5003370046615601 0.7704225182533264\n",
|
|
"500 0.4877255856990814 0.7774648070335388\n",
|
|
"600 0.47803810238838196 0.783098578453064\n",
|
|
"700 0.4705124497413635 0.8070422410964966\n",
|
|
"800 0.46460485458374023 0.8084506988525391\n",
|
|
"900 0.45992112159729004 0.8042253255844116\n",
|
|
"1000 0.456172376871109 0.8056337833404541\n",
|
|
"1100 0.45314404368400574 0.8028169274330139\n",
|
|
"1200 0.4506753087043762 0.8042253255844116\n",
|
|
"1300 0.4486445486545563 0.8014084696769714\n",
|
|
"1400 0.4469591975212097 0.7985915541648865\n",
|
|
"1500 0.4455479085445404 0.7985915541648865\n",
|
|
"1600 0.44435548782348633 0.7985915541648865\n",
|
|
"1700 0.44333896040916443 0.7957746386528015\n",
|
|
"1800 0.4424646496772766 0.797183096408844\n",
|
|
"1900 0.4417058527469635 0.7985915541648865\n",
|
|
"2000 0.44104140996932983 0.7985915541648865\n",
|
|
"2100 0.4404546320438385 0.800000011920929\n",
|
|
"2200 0.4399319589138031 0.800000011920929\n",
|
|
"2300 0.4394625723361969 0.8014084696769714\n",
|
|
"2400 0.43903765082359314 0.8014084696769714\n",
|
|
"2500 0.43865007162094116 0.8014084696769714\n",
|
|
"2600 0.4382942020893097 0.8014084696769714\n",
|
|
"2700 0.43796515464782715 0.8014084696769714\n",
|
|
"2800 0.4376590847969055 0.8014084696769714\n",
|
|
"2900 0.4373728036880493 0.8014084696769714\n",
|
|
"3000 0.4371037483215332 0.8014084696769714\n",
|
|
"3100 0.4368496239185333 0.8014084696769714\n",
|
|
"3200 0.4366086721420288 0.8014084696769714\n",
|
|
"3300 0.4363793730735779 0.8014084696769714\n",
|
|
"3400 0.43616050481796265 0.8014084696769714\n",
|
|
"3500 0.43595090508461 0.8014084696769714\n",
|
|
"3600 0.43574973940849304 0.8014084696769714\n",
|
|
"3700 0.4355562925338745 0.8014084696769714\n",
|
|
"3800 0.4353698194026947 0.8028169274330139\n",
|
|
"3900 0.4351898729801178 0.8028169274330139\n",
|
|
"4000 0.43501582741737366 0.8042253255844116\n",
|
|
"4100 0.4348474144935608 0.8042253255844116\n",
|
|
"4200 0.43468421697616577 0.8042253255844116\n",
|
|
"4300 0.4345259666442871 0.8042253255844116\n",
|
|
"4400 0.4343723654747009 0.8028169274330139\n",
|
|
"4500 0.4342231750488281 0.8014084696769714\n",
|
|
"4600 0.4340782165527344 0.8014084696769714\n",
|
|
"4700 0.4339372515678406 0.8014084696769714\n",
|
|
"4800 0.4338001012802124 0.800000011920929\n",
|
|
"4900 0.4336666464805603 0.800000011920929\n",
|
|
"5000 0.4335367679595947 0.800000011920929\n",
|
|
"5100 0.4334102272987366 0.800000011920929\n",
|
|
"5200 0.43328699469566345 0.800000011920929\n",
|
|
"5300 0.4331670105457306 0.8014084696769714\n",
|
|
"5400 0.4330500662326813 0.800000011920929\n",
|
|
"5500 0.43293607234954834 0.800000011920929\n",
|
|
"5600 0.4328249990940094 0.800000011920929\n",
|
|
"5700 0.4327167272567749 0.8028169274330139\n",
|
|
"5800 0.43261122703552246 0.8028169274330139\n",
|
|
"5900 0.432508260011673 0.8028169274330139\n",
|
|
"6000 0.432407945394516 0.8028169274330139\n",
|
|
"6100 0.43231016397476196 0.8028169274330139\n",
|
|
"Convergence reached at step: 6100\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"LR = 0.003\n",
|
|
"STEPS = 100000\n",
|
|
"\n",
|
|
"# Initial values for the parameters; requires_grad=True means that we want to compute the gradients\n",
|
|
"w = torch.zeros(6, requires_grad=True)\n",
|
|
"b = torch.zeros(1, requires_grad=True)\n",
|
|
"\n",
|
|
"# Momentum is modified to speed up the process\n",
|
|
"optimizer = torch.optim.SGD([w, b], lr=LR, momentum=0.9)\n",
|
|
"\n",
|
|
"# Arrays to store the values of the loss and accuracy for plotting\n",
|
|
"steps = []\n",
|
|
"losses = []\n",
|
|
"accuracies = []\n",
|
|
"\n",
|
|
"# Initial values for the \"precedent loss\" and \"convergence\" variables, used to check convergence\n",
|
|
"prec_loss = 0\n",
|
|
"convergence = 0\n",
|
|
"\n",
|
|
"\n",
|
|
"for step in range(STEPS):\n",
|
|
" optimizer.zero_grad()\n",
|
|
" p = logreg_inference(w, b, X)\n",
|
|
"\n",
|
|
" # Loss computation\n",
|
|
" loss = torch.nn.functional.binary_cross_entropy(p, Y.float())\n",
|
|
"\n",
|
|
" # Gradient computation and optimizer step\n",
|
|
" loss.backward()\n",
|
|
" optimizer.step()\n",
|
|
"\n",
|
|
" # Every 100 iteration the values of accuracy and loss are saved for plotting\n",
|
|
" if step % 100 == 0:\n",
|
|
" Yhat = (p > 0.5).long()\n",
|
|
" accuracy = (Yhat == Y).float().mean()\n",
|
|
"\n",
|
|
" print(step, loss.item(), accuracy.item())\n",
|
|
" steps.append(step)\n",
|
|
" losses.append(loss.item())\n",
|
|
" accuracies.append(accuracy.item())\n",
|
|
"\n",
|
|
" # Difference between \"precedent loss\" and \"current loss\"\n",
|
|
" diff = torch.absolute(prec_loss - loss)\n",
|
|
" prec_loss = loss\n",
|
|
"\n",
|
|
" #If convergence is reached, break the loop\n",
|
|
" if diff < 0.0001:\n",
|
|
" convergence = step\n",
|
|
" break\n",
|
|
"\n",
|
|
"print(\"Convergence reached at step:\", convergence)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"metadata": {
|
|
"id": "HBxbs4pQ9yn8"
|
|
},
|
|
"outputs": [
|
|
{
|
|
"data": {
|
|
"text/plain": [
|
|
"<matplotlib.legend.Legend at 0x7f6014ab3770>"
|
|
]
|
|
},
|
|
"execution_count": 12,
|
|
"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": [
|
|
"# Plots\n",
|
|
"\n",
|
|
"plt.figure()\n",
|
|
"plt.plot(steps, losses, label=\"Loss\")\n",
|
|
"plt.plot(steps, accuracies, label=\"Accuracy\")\n",
|
|
"plt.xlabel(\"Iterations\")\n",
|
|
"plt.legend()\n"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {
|
|
"id": "fW2oc5_L_bWT"
|
|
},
|
|
"source": [
|
|
"## Analyze the model\n",
|
|
"\n",
|
|
"My survival probability:"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 15,
|
|
"metadata": {
|
|
"id": "XZXWEUvS-4A2"
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"tensor([[ 2., 0., 23., 0., 2., 35.]]) torch.Size([1, 6])\n",
|
|
"tensor([0.2273], grad_fn=<SigmoidBackward0>)\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"# Personal survival probability estimation\n",
|
|
"my = torch.tensor([2, 0, 23, 0, 2, 35]).float().unsqueeze(0)\n",
|
|
"print(my, my.shape)\n",
|
|
"\n",
|
|
"\n",
|
|
"print(logreg_inference(w, b, my))"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {
|
|
"id": "REtXVhrlCvja"
|
|
},
|
|
"source": [
|
|
"Print learned weights to make considerations (in the report):"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 16,
|
|
"metadata": {
|
|
"id": "WKXx3KJNCxmK"
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"tensor([-0.9506, 2.7649, -0.0319, -0.3003, -0.0992, 0.0039],\n",
|
|
" requires_grad=True)\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"print(w)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {
|
|
"id": "ho3bj5kXCyyC"
|
|
},
|
|
"source": [
|
|
"What kind of passengers was most likely to survive? And what\n",
|
|
" kind to to die?"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 18,
|
|
"metadata": {
|
|
"id": "_BGBL3RFC5P1"
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"tensor([ 3.0000, 0.0000, 20.0000, 8.0000, 2.0000, 69.5500])\n",
|
|
"tensor([ 1.0000, 1.0000, 35.0000, 0.0000, 0.0000, 512.3292])\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"# a parte le considerazioni, posso andare a vedere che persone nel training set hanno prob maggiore e minore\n",
|
|
"p = logreg_inference(w, b, X)\n",
|
|
"print(X[p.argmin(0), :])\n",
|
|
"print(X[p.argmax(0), :])"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {
|
|
"id": "IDpVvB0IC4Py"
|
|
},
|
|
"source": [
|
|
"Scatter plot of the two most influential classes"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 23,
|
|
"metadata": {
|
|
"id": "eeNiB-b8C6Gt"
|
|
},
|
|
"outputs": [
|
|
{
|
|
"data": {
|
|
"text/plain": [
|
|
"Text(0, 0.5, 'Sex')"
|
|
]
|
|
},
|
|
"execution_count": 23,
|
|
"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": [
|
|
"noisy_X = X + torch.randn_like(X) * 0.1\n",
|
|
"\n",
|
|
"scatter = plt.scatter(noisy_X[:, 0], noisy_X[:, 1], c=Y, cmap=\"rainbow\", alpha=0.25)\n",
|
|
"plt.legend(handles=scatter.legend_elements()[0], labels=[\"Die\", \"Survive\"])\n",
|
|
"plt.xlabel(\"Ticket Class\")\n",
|
|
"plt.ylabel(\"Sex\")\n"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {
|
|
"id": "Ou_8-ilFEQo8"
|
|
},
|
|
"source": [
|
|
"## Evaluate the script\n",
|
|
"\n",
|
|
"Load the test set in the `Xtest` and `Ytest` tensors. Then, answer the questions."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 24,
|
|
"metadata": {
|
|
"id": "XCCRLSJgEbiY"
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"torch.Size([177, 6]) torch.float32\n",
|
|
"torch.Size([177]) torch.int64\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"# Load test data\n",
|
|
"f = open(\"titanic-test.txt\")\n",
|
|
"data = [float(x) for x in f.read().split()]\n",
|
|
"f.close()\n",
|
|
"\n",
|
|
"# Same shaping as training set\n",
|
|
"data = torch.tensor(data).view(-1, 7)\n",
|
|
"X_test = data[:, :6]\n",
|
|
"Y_test = data[:, 6].long()\n",
|
|
"\n",
|
|
"print(X_test.shape, X_test.dtype)\n",
|
|
"print(Y_test.shape, Y_test.dtype)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {
|
|
"id": "-ifmB9xQEd2A"
|
|
},
|
|
"source": [
|
|
"Test accuracy"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 25,
|
|
"metadata": {
|
|
"id": "ik2H0i0PEmpc"
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"0.7966101765632629\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"with torch.no_grad():\n",
|
|
" p_test = logreg_inference(w, b, X_test)\n",
|
|
" predictions = (p_test > 0.5).long()\n",
|
|
" accuracy = (predictions == Y_test).float().mean()\n",
|
|
"print(accuracy.item())"
|
|
]
|
|
}
|
|
],
|
|
"metadata": {
|
|
"colab": {
|
|
"provenance": []
|
|
},
|
|
"kernelspec": {
|
|
"display_name": "lab",
|
|
"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": "3.13.3"
|
|
}
|
|
},
|
|
"nbformat": 4,
|
|
"nbformat_minor": 0
|
|
}
|