WASDE_AI.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "cell_execution_strategy": "setup",
      "authorship_tag": "ABX9TyM1064J56nfJkPXuhcyundG",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/ben854719/WASDE-Revenue-in-2024/blob/main/WASDE_AI.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "source": [
        "from IPython import get_ipython\n",
        "from IPython.display import display\n",
        "# %%\n",
        "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.preprocessing import StandardScaler, LabelEncoder\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.optim as optim\n",
        "\n",
        "# Import the dataset from WASDE.\n",
        "WASDE_DATAE_df = pd.read_csv('WASDE_DATAE.csv')\n",
        "\n",
        "# Create Scaler Object.\n",
        "scaler = StandardScaler()\n",
        "\n",
        "# Numpy array for the dataset.\n",
        "numerical_features = ['code', 'report_month', 'region', 'commodity', 'item', 'year', 'period', 'value']\n",
        "\n",
        "for feature in numerical_features:\n",
        "    if WASDE_DATAE_df[feature].dtype == object:\n",
        "        le = LabelEncoder()\n",
        "        WASDE_DATAE_df[feature] = le.fit_transform(WASDE_DATAE_df[feature])\n",
        "\n",
        "# fit the scaler to the numerical features.\n",
        "scaler.fit(WASDE_DATAE_df[numerical_features])\n",
        "\n",
        "# Transform the data.\n",
        "WASDE_DATAE_df[numerical_features] = scaler.transform(WASDE_DATAE_df[numerical_features])\n",
        "\n",
        "# Create the tensor of input features.\n",
        "X = torch.tensor(WASDE_DATAE_df[['code', 'report_month', 'region', 'commodity', 'item', 'year', 'period']].values, dtype=torch.float32)\n",
        "Y = torch.tensor(WASDE_DATAE_df['value'].values, dtype=torch.float32)\n",
        "\n",
        "# train the model of the dataset.\n",
        "x_train, x_test, y_train, y_test = train_test_split(X, Y, test_size=0.2, random_state=42)\n",
        "\n",
        "# set a random seed.\n",
        "torch.manual_seed(42)\n",
        "\n",
        "# Scale target variable\n",
        "scaler_y = StandardScaler()\n",
        "y_train = scaler_y.fit_transform(y_train.reshape(-1, 1)).flatten()\n",
        "y_test = scaler_y.transform(y_test.reshape(-1, 1)).flatten()\n",
        "\n",
        "# Convert y_train and y_test back to PyTorch tensors and reshape\n",
        "y_train = torch.tensor(y_train, dtype=torch.float32).unsqueeze(1)\n",
        "y_test = torch.tensor(y_test, dtype=torch.float32).unsqueeze(1)\n",
        "\n",
        "\n",
        "# Check for NaN/Inf in input and target\n",
        "x_train[torch.isnan(x_train)] = 0\n",
        "x_train[torch.isinf(x_train)] = 0\n",
        "y_train[torch.isnan(y_train)] = 0\n",
        "y_train[torch.isinf(y_train)] = 0\n",
        "x_test[torch.isnan(x_test)] = 0\n",
        "x_test[torch.isinf(x_test)] = 0\n",
        "y_test[torch.isnan(y_test)] = 0\n",
        "y_test[torch.isinf(y_test)] = 0\n",
        "\n",
        "# Define the model using nn.Sequential.\n",
        "model = nn.Sequential(\n",
        "    nn.Linear(7, 32),\n",
        "    nn.ReLU(),\n",
        "    nn.Dropout(0.2),\n",
        "    nn.Linear(32, 16),\n",
        "    nn.ReLU(),\n",
        "    nn.Linear(16, 1)\n",
        ")\n",
        "\n",
        "# MSE Loss Function + optimizer.\n",
        "loss_fn = nn.MSELoss()\n",
        "optimizer = optim.Adam(model.parameters(), lr=0.001)\n",
        "\n",
        "num_epochs = 200\n",
        "\n",
        "epoch_format = 'Epoch[{0}' + str(num_epochs)+'], MSE.loss:{1}'\n",
        "\n",
        "for epoch in range(num_epochs):\n",
        "    predictions = model(x_train)\n",
        "    MSE = loss_fn(predictions, y_train)\n",
        "\n",
        "    # Check if loss is NaN and break.\n",
        "    if torch.isnan(MSE):\n",
        "        print(f\"Loss is NaN at epoch {epoch + 1}. Stopping training.\")\n",
        "        break\n",
        "\n",
        "    MSE.backward()\n",
        "\n",
        "    # Check for NaN/Inf in model parameters and break.\n",
        "    for name, param in model.named_parameters():\n",
        "        if torch.isnan(param.grad).any() or torch.isinf(param.grad).any():\n",
        "            print(f\"NaN or Inf detected in gradients for parameter '{name}' at epoch {epoch + 1}.\")\n",
        "            break\n",
        "\n",
        "    # Replace NaN/Inf in gradients with 0\n",
        "    for name, param in model.named_parameters():\n",
        "        if torch.isnan(param.grad).any() or torch.isinf(param.grad).any():\n",
        "            param.grad.data.fill_(0)\n",
        "\n",
        "  # Gradient Clipping.\n",
        "    torch.nn.utils.clip_grad_norm_(model.parameters(), max_norm=1)\n",
        "\n",
        "for m in model.modules():\n",
        "    if isinstance(m, nn.Linear):\n",
        "        torch.nn.init.xavier_uniform_(m.weight)\n",
        "        m.bias.data.fill_(0.01)\n",
        "\n",
        "    optimizer.step()\n",
        "    optimizer.zero_grad()\n",
        "\n",
        "    if (epoch+1) % 100 == 0:\n",
        "        print(epoch_format.format(epoch + 1, MSE.item()))"
      ],
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Jg7AaaShq2VJ",
        "outputId": "e4bb1fc7-c977-4e37-c022-cc47c2eb47e0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch[200200], MSE.loss:0.9669618010520935\n",
            "Epoch[200200], MSE.loss:0.9669618010520935\n",
            "Epoch[200200], MSE.loss:0.9669618010520935\n",
            "Epoch[200200], MSE.loss:0.9669618010520935\n",
            "Epoch[200200], MSE.loss:0.9669618010520935\n",
            "Epoch[200200], MSE.loss:0.9669618010520935\n",
            "Epoch[200200], MSE.loss:0.9669618010520935\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# formula for torch save.\n",
        "torch.save(model.state_dict(), 'model.pth')"
      ],
      "metadata": {
        "id": "EqHpZFfGGatt"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "source": [
        "# Load the model.\n",
        "model = nn.Sequential(\n",
        "    nn.Linear(7, 32),\n",
        "    nn.ReLU(),\n",
        "    nn.Dropout(0.2),\n",
        "    nn.Linear(32, 16),\n",
        "    nn.ReLU(),\n",
        "    nn.Linear(16, 1)\n",
        ")\n",
        "\n",
        "for m in model.modules():\n",
        "   if isinstance(m, nn.Linear):\n",
        "        torch.nn.init.xavier_uniform_(m.weight)\n",
        "        m.bias.data.fill_(0.01)  # Initialize bias to a small value\n",
        "\n",
        "# Load the state dictionary directly into the model\n",
        "model.load_state_dict(torch.load('model.pth', weights_only=True))\n",
        "\n",
        "# Set the model to evaluation mode.\n",
        "model.eval()\n",
        "\n",
        "#Replace NaN Valies in x_test with 0 (or mean/medium).\n",
        "x_test[torch.isnan(x_test)] = 0\n",
        "\n",
        "# Make predictions and calculate MSE.\n",
        "with torch.no_grad():\n",
        "    predictions = model(x_test)\n",
        "    MSE = loss_fn(predictions, y_test)\n",
        "    print(MSE)"
      ],
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "exZeb76drG90",
        "outputId": "77721fc6-a700-4dff-9187-b62ad39825f6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "tensor(1.3846)\n"
          ]
        }
      ]
    },
    {
      "source": [
        "# Data visualization for the dataset.\n",
        "\n",
        "# Group data by region and get the values for each region.\n",
        "region_values = WASDE_DATAE_df.groupby('region')['value'].apply(list).reset_index()\n",
        "\n",
        "# Create the histogram.\n",
        "plt.figure(figsize=(10, 6))\n",
        "for index, row in region_values.iterrows():\n",
        "    plt.hist(row['value'], bins=10, alpha=0.7, label=row['region'], edgecolor='black')\n",
        "\n",
        "plt.xlabel(\"Value\")\n",
        "plt.ylabel(\"Frequency\")\n",
        "plt.title(\"Distribution of Value by Region\")\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 564
        },
        "id": "VqSR8RbB85M4",
        "outputId": "b57d40bf-0f77-4974-f1a6-f0f511bae94f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "Explanation:\n",
        "\n",
        "The histogram graph explains the frequency and the value of commodities of good in the different region of the world. You will notice the bar one has the lowest frequency of 0.99 and the lowest value compared to the other graph. The graph 1 represents the lowest of value of trade of commodities.\n",
        "You will notice that the bar 2 and 3 that overshadowed each other as the high frequency of 5.9 which the revenue per capita of trade of goods compare to the regions of the world.The bar 4, 5, and 6 have the same frequency of 3.8 which means their values of trade in revenue for goods are the same in their categories.\n",
        "\n"
      ],
      "metadata": {
        "id": "XI__G_3A9WU0"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}