{ "cells": [ { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "# Causal-Retro-Causal Neural Network\n", "\n", "This Notebook presents the Causal-Retro-Causal Neural Network (CRCNN). Because the CRCNN is a further development of Historical Consistant Neural Networks, we assume that you are already familiar with its [tutorial](HCNN.ipynb). Based on this we will look into the theory and the ideas of the CRCNN architecture. Then the data is prepared for model training. There are some ways presented how to set up a training loop. We start with the simple case without the mirroring trick, which is then introduced in the following section. We will use the mirroring trick and train a model at first only on one batch and second with multiple batches. At the end of the notebook we will build an ensemble of CRCNNs." ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "## Theory\n", "\n", "As mentioned, the CRCNN is a further development of the HCNN. The HCNN is restricted that only influences of the past can affect the future. In reality, decisions in the present are often made with respect to interests in the future. For instance, you invest into stocks because you believe it has an increasing value in the future. This decision is made in the present with motivation coming from the future. To represent the influence of the future into the past we change the direction of time in a HCNN (retro-causal model). \n", "We combine both ideas of causal (normal HCNN) and retro-causal (inverted HCNN) in one model. The resulting CRCNNs predictions are a combination of both influences (see figure below). \n", "\n", "\n", "\n", "Because of the teacher forcing, we have closed loops (orange) in the architecture. Since the backpropagation algorithm propagates the gradients against the direction of the arrow and therefore in loops, the algorithm does not converge. But losing the teacher forcing in order to get rid of the loops would give us an untrainable model. A suited solution is to approximate the above architecture with a finite unfolding of the loop.\n", "\n", "In its simplest version we copy the causal branch of the model and stack it on top of the retro-causal branch (picture below). To get rid of the loop, we delete the teacher forcing in the first causal branch. This is possible because of the shared weights to the top causal branch, where the teacher forcing is applied. This way the first branch is also trainable. \n", "As a result, both the causal and the retro-causal branch impact the outputs. There are now two versions of outputs. On the one hand, there is the output between the first causal and the retro-causal branch, but there is no teacher forcing in the first branch. On the other hand, we have the top output between the top causal and the retro-causal branch. This output is a combination of two branches with teacher forcing and therefore gives better results.\n", "\n", "However, since the possible worse predictions in the first output, the teacher forcing inserts poor information from the causal branch into the retro-causal branch. This continues to the top output. To mitigate this we add alternating copies of causal and retro-causal branches. In each new branch the problem decreases until we have an output in the highest branch that is only slightly affected by the missing teacher forcing in the first branch. The number of branches is under research, but in most cases it is enough to have around 7 branches.\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(\".\"))" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 2, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import torch\n", "\n", "from prosper_nn.models.crcnn import CRCNN\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", "from prosper_nn.utils import visualize_forecasts\n", "\n", "torch.manual_seed(0)" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "## Data Preparation\n", "\n", "The data preparation is similar to that of the HCNN, which you can see in its [tutorial](Hcnn.ipynb#data-preparation)." ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "past_horizon = 30\n", "forecast_horizon = 5\n", "n_features_Y = 2\n", "batchsize = 1\n", "n_batches = 2\n", "n_data = 100" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "# generate data with \"unknown\" variables U\n", "Y, U = gtsd.sample_data(n_data, n_features_Y=n_features_Y - 1, n_features_U=1)\n", "Y = torch.cat((Y, U), 1)\n", "\n", "# Only use Y as input for the hcnn\n", "Y_batches = ci.create_input(\n", " Y=Y,\n", " past_horizon=past_horizon,\n", " batchsize=batchsize,\n", " forecast_horizon=forecast_horizon,\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Similar to the HCNN, the outputs of the model are trained to be zero in the `past_horizon`." ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "targets_past = torch.zeros((past_horizon, batchsize, n_features_Y))" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "## Single Causal-Retro-Causal Neural Network (CRCNN)\n", "\n", "The chapter gives an overview over the CRCNN model without the mirroring trick. We start with the initialization, continue with the training and finish with forecasting a time series." ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Initialization\n", "\n", "The initialization is nearly identical to that of the HCNN because it is implemented with the help of the [HCNNCell](../api/hcnn.rst#prosper_nn.models.hcnn.hcnn_cell.HCNNCell). Therefore, the parameters for initializing the HCNN are also necessary for the CRCNN. But we have an additional `n_branches` parameter. It is the total sum of causal and retro-causal branches in the model. The CRCNN is implemented in such a way that the first branch is always a causal branch. The second will be a retro-causal branch. While increasing the number, copys of the causal and retro-causal branches are added to the model in turns. \n", "The second additional parameter is the `batchsize` to be able to initialize the `future_bias` that is introduced in the following section." ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": [ "n_state_neurons = 20\n", "n_branches = 5" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [], "source": [ "crcnn_model = CRCNN(\n", " n_state_neurons=n_state_neurons,\n", " n_features_Y=n_features_Y,\n", " past_horizon=past_horizon,\n", " forecast_horizon=forecast_horizon,\n", " n_branches=n_branches,\n", " batchsize=batchsize,\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Set the `optimizer` and the `loss_function`." ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [], "source": [ "optimizer = torch.optim.Adam(crcnn_model.parameters())\n", "loss_function = torch.nn.MSELoss()" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Training Loop\n", "\n", "The training loop is designed analogous to the HCNN. The only change is that the loss now contains the errors of all outputs in the `past_horizon`. Because there is always an output between a causal and a retro-causal branch, there are `(n_branches - 1) * past_horizon` output layers relevant." ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [], "source": [ "epochs = 10\n", "\n", "total_loss = epochs * [0]\n", "for epoch in range(epochs):\n", " for batch_index in range(0, Y_batches.shape[0]):\n", " crcnn_model.zero_grad()\n", "\n", " Y_batch = Y_batches[batch_index]\n", " model_output = crcnn_model(Y_batch)\n", " past_errors, forecasts = torch.split(model_output, past_horizon, dim=1)\n", "\n", " losses_past = [\n", " loss_function(past_errors[i][j], targets_past[j])\n", " for i in range(n_branches - 1)\n", " for j in range(past_horizon)\n", " ]\n", " loss = sum(losses_past) / len(losses_past)\n", " loss.backward()\n", " optimizer.step()\n", " total_loss[epoch] += loss.detach()" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Forecast\n", "\n", "Again, forecasts can be calculated similarly as with the HCNN model. Theoretically the best prediction is in the top output. So, we choose the last output in `forecasts` as our final `forecast`." ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [], "source": [ "with torch.no_grad():\n", " crcnn_model.eval()\n", "\n", " output_forecast = crcnn_model(Y_batches[0, :, 0].unsqueeze(1))\n", " past_errors, forecast = torch.split(output_forecast[-1], past_horizon)\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Evaluation\n", "\n", "We want to visualize the predicted outcome of the model." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Postprocessing\n", "\n", "Because of the different meaning of the output along the `past_horizon` compared to the output along the `forecast_horizon`, we have to calculate the `expected_timeseries` first before we can visualize the expectation/target comparison." ] }, { "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", ")" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "## Causal-Retro-Causaul Neural Network with Mirroring\n", "\n", "This chapter will proceed as the chapter before, but the mirror trick is used. \n", "Our goal is to have a symmetric model in the past and future. But when we look into the architecture of the CRCNN, the teacher forcing is only applied during the past. This loop induces equations in the architecture, since after one loop the values have to be the same as before. This can be interpreted as training on a manifold. In the future we don't have the observations to compare the output with the target. Therefore, it is possible for the model to leave the manifold. To stay on it, we use a fake observation in the future. Then we can include teacher forcing in the future, which forces the model on the manifold. \n", "\n", "To get the fake observations we see the forecasts as trainable parameters $\\tilde{y}_{t+1}^d, \\dots \\tilde{y}_{t+n}^d$, like biases, in the output layers. Therefore, $\\tilde{y}^d$ will be named `future_bias` in the code. On the one hand, this bias works as an observation in the future and on the other hand, it is the forecast of the model. In this way we get a symmetric model in past and future. \n", "\n", "\n", "\n", "The drawback is that each $\\tilde{y}^d$ has to be implemented as a learnable parameter. Therefore, it is not simply possible to use different batches during the training and we have to use all data in one batch. Nevertheless, we will discuss a possible solution later." ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Initialization\n", "\n", "Compared to the CRCNN without mirroring, we only set the parameter `mirroring=True` and the mirroring trick is applied." ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [], "source": [ "mirroring = True" ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [], "source": [ "crcnn_future_bias_model = CRCNN(\n", " n_state_neurons=n_state_neurons,\n", " n_features_Y=n_features_Y,\n", " past_horizon=past_horizon,\n", " forecast_horizon=forecast_horizon,\n", " n_branches=n_branches,\n", " batchsize=batchsize,\n", " mirroring=mirroring,\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Additionally, we need targets for the fake outputs in the future. The objective is to have, like in the past, zero loss since $\\tilde{y}^d$ works as observation. So instead of the direct forecast as it was in the previous model we get the `forecast_errors`. This results from the `future_bias` being subtracted from the sum of causal and retro-causal states: $forecast\\_errors_t = s_t + s'_t - future\\_bias_t$." ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [], "source": [ "targets_future = torch.zeros((forecast_horizon, batchsize, n_features_Y))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Reset the optimizer." ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [], "source": [ "optimizer = torch.optim.Adam(crcnn_future_bias_model.parameters())" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Training Loop\n", "\n", "Because the fake observations have to be learned by the model, it is only possible to learn with data in one batch. We will just use the first batch.\n", "\n", "For the backpropagation, it is now important to add the `forecast_errors` to the calculation of the loss." ] }, { "cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [], "source": [ "epochs = 10\n", "total_loss = epochs * [0]\n", "Y_batch = Y_batches[0]\n", "for epoch in range(epochs):\n", " crcnn_future_bias_model.zero_grad()\n", "\n", " model_output = crcnn_future_bias_model(Y_batch)\n", " past_errors, forecast_errors = torch.split(model_output, past_horizon, dim=1)\n", "\n", " losses_past = [\n", " loss_function(past_errors[i][j], targets_past[j])\n", " for i in range(n_branches - 1)\n", " for j in range(past_horizon)\n", " ]\n", " losses_mirror = [\n", " loss_function(forecast_errors[i][j], targets_future[j])\n", " for i in range(n_branches - 1)\n", " for j in range(forecast_horizon)\n", " ]\n", " loss = sum(losses_mirror) / len(losses_mirror) + sum(losses_past) / len(losses_past)\n", " loss.backward()\n", " optimizer.step()\n", " total_loss[epoch] += loss.detach()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Forecasts\n", "\n", "In the forecast it is not necessary anymore to use the mirroring, since the model is trained to zero loss. Therefore, the teacher forcing vector is approximately zero and we can leave it out of the model. So, we set `mirroring=False` before calculating the forecast.\n", "\n", "Although the teacher forcing is close to zero, the results are a little bit different in each output branch because in the first branch no teacher forcing is applied. We expect the most reliable result between last causal and retro-causal branches." ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [], "source": [ "with torch.no_grad():\n", " crcnn_future_bias_model.eval()\n", " crcnn_future_bias_model.mirroring = False\n", "\n", " output_forecast = crcnn_future_bias_model(Y_batches[0, :, 0].unsqueeze(1))\n", " past_errors, forecast = torch.split(output_forecast[-1], past_horizon)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Evaluation\n", "\n", "We want to compare the predicted output with the real data." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Postprocessing\n", "\n", "Before we can plot the time series, we have to postprocess the predictions of the model. Like in the HCNN we have to add the true data on the predictions during the `past_horizon`." ] }, { "cell_type": "code", "execution_count": 19, "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", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Causal-Retro-Causal Neural Network with Mirroring and Multiple Batches\n", "\n", "Because the forecasts of the model with mirroring are learnable parameters, it is not possible to simply change data or batches during training and get reasonable predictions. But if we have too many data for one batch, we can design the training loop a little bit differently to overcome this limitation. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Initialization\n", "\n", "We want to use the data with different batches again. Therefore we reset the model." ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [], "source": [ "crcnn_future_bias_model = CRCNN(\n", " n_state_neurons=n_state_neurons,\n", " n_features_Y=n_features_Y,\n", " past_horizon=past_horizon,\n", " forecast_horizon=forecast_horizon,\n", " n_branches=n_branches,\n", " batchsize=batchsize,\n", " mirroring=mirroring,\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Until now we always used the Adam optimizer. But in this example it is possibly not a good idea to use it. Adam has a momentum term when it applies an optimizer step. When we switch from one batch to another the momentum of the previous batch affects the recent batch. For learning the matrices this is a good idea but for learning the `future_bias` probably not. It is possible that this isn't problematic because the objective to learn the forecast is very easy (linear). Nevertheless, we recommend to use an optimizer without momentum or reset the optimizer between different batches." ] }, { "cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [], "source": [ "optimizer = torch.optim.SGD(crcnn_future_bias_model.parameters(), lr=0.001)" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Training Loop\n", "\n", "As already mentioned, the training loop has to be modified to train on multiple batches. The problem is the changing fake observation in the future. With only one batch it is possible to train the `future_bias` during training of the model. With multiple batches this isn't possible because the observations are changing and with them the predictions. To have the right fake observation in the model we save the `future_bias` in a tensor `store_bias` and initialize the data parameter of the `future_bias` in each batch with the according value. Now, we can train the model as in the previous section." ] }, { "cell_type": "code", "execution_count": 22, "metadata": {}, "outputs": [], "source": [ "store_bias = torch.empty(\n", " (Y_batches.shape[0], forecast_horizon, batchsize, n_features_Y)\n", ")\n", "\n", "epochs = 10\n", "total_loss = epochs * [0]\n", "for epoch in range(epochs):\n", " for batch_index in range(0, Y_batches.shape[0]):\n", " crcnn_future_bias_model.zero_grad()\n", "\n", " Y_batch = Y_batches[batch_index]\n", " # set the future_bias from store_bias accordingly to the batch\n", " if epoch == 0:\n", " crcnn_future_bias_model.future_bias.data = torch.zeros(\n", " size=(forecast_horizon, batchsize, n_features_Y)\n", " )\n", " else:\n", " crcnn_future_bias_model.future_bias.data = store_bias[batch_index]\n", " model_output = crcnn_future_bias_model(Y_batch)\n", " past_error, forecast = torch.split(model_output, past_horizon, dim=1)\n", "\n", " losses_past = [\n", " loss_function(past_error[i][j], targets_past[j])\n", " for i in range(n_branches - 1)\n", " for j in range(past_horizon)\n", " ]\n", " losses_mirror = [\n", " loss_function(forecast[i][j], targets_future[j])\n", " for i in range(n_branches - 1)\n", " for j in range(forecast_horizon)\n", " ]\n", " loss = sum(losses_mirror) / len(losses_mirror) + sum(losses_past) / len(\n", " losses_past\n", " )\n", " loss.backward()\n", " optimizer.step()\n", " total_loss[epoch] += loss.detach()\n", " # store the new bias in store_bias\n", " store_bias[batch_index] = crcnn_future_bias_model.future_bias.data.detach()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Forecast\n", "\n", "The forecast is calculated like for one batch. We will make a forecast for the first batch in the data set." ] }, { "cell_type": "code", "execution_count": 23, "metadata": {}, "outputs": [], "source": [ "with torch.no_grad():\n", " crcnn_future_bias_model.eval()\n", " crcnn_future_bias_model.mirroring = False\n", "\n", " output_forecast = crcnn_future_bias_model(Y_batches[0, :, 0].unsqueeze(1))\n", " past_errors, forecast = torch.split(output_forecast[-1], past_horizon)\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Evaluation\n", "\n", "Also the evaluation is identical to that of the one batch version." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Postprocessing\n", "\n", "Again the true data is added to the `past_errors` and than concatenated to the forecast. " ] }, { "cell_type": "code", "execution_count": 24, "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", ")" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "## Ensemble of Causal-Retro-Causal Neural Networks\n", "\n", "In addition to the CRCNN being a very complex model, the forecasts are still uncertain due to over-parameterization and random initialization. We get rid of this with building an ensemble and using the median forecast. This section builds an ensemble of an CRCNN without mirroring." ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Initialization\n", "\n", "The ensemble is built the same way as an ensemble of an HCNN (see [here](Hcnn.ipynb#Ensemble-of-Historical-Consistent-Neural-Networks)). First we have to reset the model itself and then we can use the `Ensemble` class to calculate multiple models at once." ] }, { "cell_type": "code", "execution_count": 25, "metadata": {}, "outputs": [], "source": [ "n_models = 3" ] }, { "cell_type": "code", "execution_count": 26, "metadata": {}, "outputs": [], "source": [ "crcnn_model = CRCNN(\n", " n_state_neurons=n_state_neurons,\n", " n_features_Y=n_features_Y,\n", " past_horizon=past_horizon,\n", " forecast_horizon=forecast_horizon,\n", " n_branches=n_branches,\n", " batchsize=batchsize,\n", ")" ] }, { "cell_type": "code", "execution_count": 27, "metadata": {}, "outputs": [], "source": [ "crcnn_ensemble = Ensemble(\n", " model=crcnn_model, n_models=n_models, combination_type=\"median\"\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Reset the optimizer. Because we don't use mirroring here, we can go back to the Adam optimizer." ] }, { "cell_type": "code", "execution_count": 28, "metadata": {}, "outputs": [], "source": [ "optimizer = torch.optim.Adam(crcnn_ensemble.parameters())" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Training Loop\n", "The number of dimensions of the model output increases and we have `ensemble_output.shape() = (n_models + 1, n_branches - 1, past_horizon + forecast_horizon, batchsize, n_features_Y)`. We use the first three dimensions to calculate the loss." ] }, { "cell_type": "code", "execution_count": 29, "metadata": {}, "outputs": [], "source": [ "epochs = 10\n", "\n", "total_loss = epochs * [0]\n", "for epoch in range(epochs):\n", " for batch_index in range(0, Y_batches.shape[0]):\n", " crcnn_ensemble.zero_grad()\n", "\n", " Y_batch = Y_batches[batch_index]\n", " ensemble_output = crcnn_ensemble(Y_batch)\n", " outputs, mean = torch.split(ensemble_output, n_models)\n", " mean = torch.squeeze(mean, 0)\n", " past_errors, forecasts = torch.split(outputs, past_horizon, dim=2)\n", "\n", " losses_past = [\n", " loss_function(past_errors[k][i][j], targets_past[j])\n", " for k in range(n_models)\n", " for i in range(n_branches - 1)\n", " for j in range(past_horizon)\n", " ]\n", " loss = sum(losses_past) / len(losses_past)\n", " loss.backward()\n", "\n", " optimizer.step()\n", "\n", " mean_loss = (\n", " sum(\n", " [\n", " loss_function(mean[-1, i], targets_past[i])\n", " for i in range(past_horizon)\n", " ]\n", " )\n", " / past_horizon\n", " )\n", " total_loss[epoch] += mean_loss.detach()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Forecast\n", "The forecast can be done similarly to the CRCNN case without mirroring. We just have to select the last entry of the ensemble dimension additionally." ] }, { "cell_type": "code", "execution_count": 30, "metadata": {}, "outputs": [], "source": [ "with torch.no_grad():\n", " crcnn_ensemble.eval()\n", "\n", " output_forecast = crcnn_ensemble(Y_batches[0, :, 0].unsqueeze(1))\n", " past_errors, forecast = torch.split(output_forecast[-1, -1], past_horizon)\n" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Evaluation\n", "\n", "To get the real time series we have to do a post-processing step.\n", "Further evaluation can be done like in the [tutorial of the HCNN](Hcnn.ipynb#id3).\n", "\n", "#### Postprocessing\n", "\n", "Again, during the past the model outputs are added to the observations and the are concatenated to the forecasts." ] }, { "cell_type": "code", "execution_count": 31, "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 }