{
"cells": [
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"# Historical Consistent Neural Network with known features\n",
"\n",
"In this notebook we demonstrate how the prosper_nn package can be used to build and analyze a Historical Consistent Neural Network (HCNN) with (partly) known features.\n",
"Similar to the other tutorials, it begins with a simple version of an HCNN_known_u and shows how the data and the training loop should look like.\n",
"If you want to build an ensemble please refer to [HCNN tutorial](Hcnn.ipynb#Ensemble-of-Historical-Consistent-Neural-Networks)."
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"## Theory\n",
"\n",
"Historical Consistent Neural Networks with known features U are based on the architectures of HCNNs and ECNNs. As a result it belongs to the class of Recurrent Neural Networks. The picture below shows the architecture of the model.\n",
"\n",
"The common HCNN model treats all features as targets and forecasts them into the future. The version of the HCNN presented in this notebook, however, distinguishes between features we want to forecast and features we already know for the future (like e.g. holidays or months). The latter is what 'known U' refers to. As a result, we use all the advantages of the common HCNN while also utilizing our known information to capacity. Instead of trying to forecast holidays in the HCNN (and failing), we provide the model with this information even along the forecast horizon. In a way, this model is like a combination of HCNN (model features internally if we don't know them for the future) and ECNN (supply the model with external features if we know them for the future).\n",
"\n",
"We facilitate the inclusion of known features by concatenating them to the state $r$. These known features are used for calculating the next state but are themselves not modeled as part of the next state. This is why $dim(r_t) = (batchsize, n\\_state\\_neurons + n\\_features\\_U)$ whereas $dim(s_t) = (batchsize, n\\_state\\_neurons)$. Accordingly, $dim(A) = ((n\\_state\\_neurons + n\\_features\\_U) , n\\_state\\_neurons)$.\n",
"\n",
"To calculate the state of the next time step, a non-linearity ($\\tanh$) and the state transition matrix $A$ are applied. These update steps describe the implementation of the HCNN_known_u cell that calculates the output and the following state for one time step. In formula each cell performs the following calculation.\n",
"For readability we use these abbreviations:\n",
" - nst = n_state_neurons \n",
" - nfY = n_features_Y \n",
" - nfU = n_features_U \n",
"\n",
"$$\\hat{z}_t = [\\mathbb{I}_{nfY}, \\mathbb{0}_{nst-nfY}] s_t -y_t^d$$\n",
"$$r_t = [\\mathbb{I}_{nst}, \\mathbb{0}_{nst, nfU}]^T \\cdot s_t - [\\mathbb{I}_{nfY}, \\mathbb{0}_{nfY, nst-nfY}, \\mathbb{0}_{nfY, nfU}]^T \\cdot \\hat{z}_t + [\\mathbb{0}_{nfU, nst}, \\mathbb{I}_{nfU}]^T u_{t+1} $$\n",
"$$ \n",
"s_{t+1} = A \\tanh (r_t) $$\n",
"$$ y_t = [\\mathbb{I}_{nfY}, \\mathbb{0}_{nst-nfY}] \\cdot s_t $$\n",
"The first part of $r$ contains the data, the middle portion the hidden features and the end the info from the external features U.\n",
"\n",
"\n",
""
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {
"tags": [
"hide_cell"
]
},
"outputs": [],
"source": [
"import sys, os\n",
"\n",
"sys.path.append(os.path.abspath(\"../../..\"))\n",
"sys.path.append(os.path.abspath(\"../..\"))\n",
"sys.path.append(os.path.abspath(\"..\"))\n",
"sys.path.append(os.path.abspath(\".\"))"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
""
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import torch\n",
"import torch.nn as nn\n",
"import torch.optim as optim\n",
"\n",
"from prosper_nn.models.hcnn_known_u import hcnn_known_u\n",
"from prosper_nn.models.ensemble import Ensemble\n",
"\n",
"import prosper_nn.utils.generate_time_series_data as gtsd\n",
"import prosper_nn.utils.create_input_ecnn_hcnn as ci\n",
"\n",
"import prosper_nn.utils.neuron_correlation_hidden_layers as nchl\n",
"import prosper_nn.utils.visualization as visualization\n",
"from prosper_nn.utils import visualize_forecasts\n",
"from prosper_nn.utils import sensitivity_analysis\n",
"torch.manual_seed(0)"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"## Data Preparation \n",
"\n",
"For the data creation look at the [ECNN tutorial](ECNN.ipynb#Data-preparation).\n",
"If we set `n_features_U` to zero, we would revert this architecture back to vanilla HCNN. "
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"past_horizon = 30\n",
"forecast_horizon = 5\n",
"n_features_U = 2\n",
"n_features_Y = 3\n",
"future_U = True\n",
"batchsize = 5\n",
"n_batches = 4\n",
"n_data = 100"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"For HCNN_known_u the target data should be in the `shape=(past_horizon, batchsize, n_features_Y)` and the external features which are known for the future in the shape `shape=(past_horizon + forecast_horizon, batchsize, n_features_U)`. "
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"c:\\Users\\bkn\\Projekte\\Prosper\\prosper\\prosper_nn\\utils\\create_input_ecnn_hcnn.py:57: UserWarning: For the last values of Y there are not enough future Us, so they will be discarded.\n",
" warnings.warn(\"For the last values of Y there are not enough \"\n",
"C:\\Users\\bkn\\AppData\\Local\\Temp\\ipykernel_19076\\3361968738.py:3: UserWarning: The number of sequences generated from the data are not a multiple of batchsize. The first 1 sequences will be discarded.\n",
" Y_batches, U_batches = ci.create_input(\n"
]
}
],
"source": [
"# generate data with n_features_U and n_features_Y\n",
"Y, U = gtsd.sample_data(n_data, n_features_Y, n_features_U)\n",
"Y_batches, U_batches = ci.create_input(\n",
" Y, past_horizon, batchsize, U, future_U, forecast_horizon\n",
")"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"Equal to HCNN, the targets of the HCNN_known_u should be in the same shape as $Y$, that is `shape=(past_horizon, batchsize, n_features_Y)`. Because the output of the HCNN_known_u is already the comparison between observation and expectation in the past horizon, the targets are zeros."
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
"targets = torch.zeros((past_horizon, batchsize, n_features_Y))"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"## Single Historical Consistent Neural Network with known features (HCNN_known_u) \n",
"\n",
"In this section, we apply a HCNN with known features. We first start with the initialization of the model. Then we discuss the training loop and create forecasts for the data we generated. At the end of the section we evaluate the model and analyze it."
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"### Initialization\n",
"\n",
"Compared to the [HCNN](../api/hcnn.rst) we only have to specify the `n_featues_U` additionally."
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [],
"source": [
"n_state_neurons = 30\n",
"sparsity = 0"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"hcnn_known_u_model = hcnn_known_u.HCNN_KNOWN_U(\n",
" n_state_neurons,\n",
" n_features_U,\n",
" n_features_Y,\n",
" past_horizon,\n",
" forecast_horizon,\n",
" sparsity,\n",
")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Set optimizer and loss function."
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [],
"source": [
"optimizer = optim.Adam(hcnn_known_u_model.parameters(), lr=0.001)\n",
"loss_function = nn.MSELoss()"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"### Training Loop\n",
"\n",
"In this training loop the output of the HCNN_known_u has `shape=(past_horizon + forecast_horizon, batchsize, n_features_Y)`. So there is no difference to the basic HCNN and we only have to give the `U_batch` in the forward of the model."
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [],
"source": [
"epochs = 150\n",
"\n",
"total_loss = epochs * [0]\n",
"for epoch in range(epochs):\n",
" for batch_index in range(0, U_batches.shape[0]):\n",
" hcnn_known_u_model.zero_grad()\n",
"\n",
" U_batch = U_batches[batch_index]\n",
" Y_batch = Y_batches[batch_index]\n",
"\n",
" model_out = hcnn_known_u_model(U_batch, Y_batch)\n",
"\n",
" past_error, forecast = torch.split(model_out, past_horizon)\n",
"\n",
" losses = [loss_function(past_error[i], targets[i]) for i in range(past_horizon)]\n",
" loss = sum(losses)\n",
" loss.backward()\n",
" \n",
" optimizer.step()\n",
" total_loss[epoch] += loss.detach()\n"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"### Forecast \n",
"\n",
"For a final prediction we only need to forward the U and Y we want to use for the actual forecast through the model. Again, we get the forecast and also the error the model still makes on the known Y."
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [],
"source": [
"example_pred_U = torch.reshape(\n",
" U[0 : (past_horizon + forecast_horizon), :],\n",
" (past_horizon + forecast_horizon, 1, n_features_U),\n",
").float()\n",
"example_pred_Y = torch.reshape(\n",
" Y[0 : (past_horizon + forecast_horizon), :],\n",
" (past_horizon + forecast_horizon, 1, n_features_Y),\n",
").float()\n",
"\n",
"with torch.no_grad():\n",
" hcnn_known_u_model.eval()\n",
"\n",
" model_output = hcnn_known_u_model(example_pred_U, example_pred_Y[:past_horizon])\n",
" past_errors, forecast = torch.split(model_output, past_horizon)\n"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"### Evaluation\n",
"#### Postprocessing\n",
"Because the output of the model has different meaning for `past_horizon` and `forecast_horizon`, the `expected_timeseries` can be calculated by adding the real observation data `Y` on the `past_error` for the `past_horizon` and concatenate the result to the `forecast` of the model."
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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",
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"expected_timeseries = torch.cat(\n",
" (torch.add(past_errors.squeeze(), Y[:past_horizon]), forecast.squeeze()), dim=0\n",
").detach()\n",
"\n",
"visualize_forecasts.plot_time_series(\n",
" expected_time_series=expected_timeseries[:, 0],\n",
" target=Y[: past_horizon + forecast_horizon, 0],\n",
")"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "prosper",
"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": "3.8.5"
},
"vscode": {
"interpreter": {
"hash": "a604604040b0261c277bc75aa34f15c6f86bb9bc8166d3b0f73ab3af3d1b81ef"
}
}
},
"nbformat": 4,
"nbformat_minor": 2
}