Add Monte Carlo prediction technique (#846)

* Add monte carlo forecast

* Add hugo

* Update _index.md

* review edits

Co-authored-by: James Maslek <james@macbook-pro.myfiosgateway.com>

gamestonk_terminal/common/prediction_techniques/mc_model.py

@@ -0,0 +1,47 @@
+"""Monte Carlo Model"""
+__docformat__ = "numpy"
+
+from typing import Union
+import pandas as pd
+import numpy as np
+from scipy.stats import norm
+
+
+def get_mc_brownian(
+    data: Union[pd.Series, np.array], n_future: int, n_sims: int, use_log=True
+) -> np.array:
+    """Performs monte carlo forecasting for brownian motion with drift
+
+    Parameters
+    ----------
+    data : Union[pd.Series, np.array]
+        Input data.
+    n_future : int
+        Number of future steps
+    n_sims : int
+        Number of simulations to run
+    use_log : bool, optional
+        Flag to use log returns, by default True
+
+    Returns
+    -------
+    np.array
+        Array of predictions.  Has shape (n_future, n_sims)
+    """
+    changes = data.pct_change().dropna()
+    if use_log:
+        changes = np.log(1 + changes)
+
+    dist_mean = changes.mean()
+    dist_var = changes.var()
+    dist_drift = dist_mean - 0.5 * dist_var
+    dist_std = np.sqrt(dist_var)
+
+    random_steps = norm.ppf(np.random.rand(n_future, n_sims))
+    predicted_change = np.exp(dist_drift + dist_std * random_steps)
+    possible_paths = np.zeros_like(predicted_change)
+    possible_paths[0] = data.values[-1]
+
+    for t in range(1, n_future):
+        possible_paths[t] = possible_paths[t - 1] * predicted_change[t]
+    return possible_paths

gamestonk_terminal/common/prediction_techniques/mc_view.py

@@ -0,0 +1,70 @@
+"""Monte Carlo View"""
+__docformat__ = "numpy"
+
+from typing import Union
+import os
+
+import pandas as pd
+import numpy as np
+import seaborn as sns
+import matplotlib.pyplot as plt
+import matplotlib.dates as mdates
+
+from gamestonk_terminal.common.prediction_techniques import mc_model
+from gamestonk_terminal.helper_funcs import (
+    plot_autoscale,
+    get_next_stock_market_days,
+    export_data,
+)
+from gamestonk_terminal.config_plot import PLOT_DPI
+from gamestonk_terminal import feature_flags as gtff
+
+
+def display_mc_forecast(
+    data: Union[pd.Series, np.array],
+    n_future: int,
+    n_sims: int,
+    use_log=True,
+    export: str = "",
+):
+    """Display monte carlo forecasting
+
+    Parameters
+    ----------
+    data : Union[pd.Series, np.array]
+        Data to forecast
+    n_future : int
+        Number of days to forecast
+    n_sims : int
+        Number of simulations to run
+    use_log : bool, optional
+        Flag to use lognormal, by default True
+    export: str
+        Format to export data
+    """
+    predicted_values = mc_model.get_mc_brownian(data, n_future, n_sims, use_log)
+    future_index = get_next_stock_market_days(data.index[-1], n_next_days=n_future)
+    dateFmt = mdates.DateFormatter("%m/%d/%Y")
+
+    fig, ax = plt.subplots(1, 2, figsize=plot_autoscale(), dpi=PLOT_DPI)
+
+    ax[0].plot(data)
+    ax[0].plot(future_index, predicted_values, alpha=0.3)
+    ax[0].set_title("Data Predictions")
+    ax[0].xaxis.set_major_formatter(dateFmt)
+    ax[0].tick_params(axis="x", labelrotation=45)
+    ax[0].grid("on")
+
+    sns.histplot(predicted_values[-1, :], ax=ax[1], kde=True)
+    ax[1].set_xlabel("Price")
+    ax[1].set_title(f"Distribution of data after {n_future} days.")
+    ax[1].set_xlim(np.min(predicted_values[-1, :]), np.max(predicted_values[-1, :]))
+    ax[1].grid("on")
+
+    fig.tight_layout(pad=2)
+
+    if gtff.USE_ION:
+        plt.ion()
+    plt.show()
+    export_data(export, os.path.dirname(os.path.abspath(__file__)), "mc")
+    print("")

gamestonk_terminal/stocks/prediction_techniques/pred_controller.py

@@ -27,6 +27,7 @@ from gamestonk_terminal.common.prediction_techniques import (
     neural_networks_view,
     regression_view,
     pred_helper,
+    mc_view,
 )
 from gamestonk_terminal.stocks.stocks_helper import load
 
@@ -46,6 +47,7 @@ class PredictionTechniquesController:
         "rnn",
         "lstm",
         "conv1d",
+        "mc",
     ]
     CHOICES += CHOICES_MODELS
 
@@ -105,6 +107,7 @@ Models:
     rnn         Recurrent Neural Network
     lstm        Long-Short Term Memory
     conv1d      1D Convolutional Neural Network
+    mc          Monte-Carlo simulations
         """
         print(help_string)
 
@@ -724,6 +727,59 @@ Models:
         finally:
             pred_helper.restore_env()
 
+    def call_mc(self, other_args: List[str]):
+        """Process mc command"""
+        parser = argparse.ArgumentParser(
+            formatter_class=argparse.ArgumentDefaultsHelpFormatter,
+            add_help=False,
+            prog="mc",
+            description="""
+                Perform Monte Carlo forecasting
+            """,
+        )
+        parser.add_argument(
+            "-d",
+            "--days",
+            help="Days to predict",
+            dest="n_days",
+            type=check_positive,
+            default=30,
+        )
+        parser.add_argument(
+            "-n",
+            "--num",
+            help="Number of simulations to perform",
+            dest="n_sims",
+            default=100,
+        )
+        parser.add_argument(
+            "--dist",
+            choices=["normal", "lognormal"],
+            default="lognormal",
+            dest="dist",
+            help="Whether to model returns or log returns",
+        )
+
+        try:
+            ns_parser = parse_known_args_and_warn(
+                parser, other_args, export_allowed=EXPORT_ONLY_FIGURES_ALLOWED
+            )
+            if not ns_parser:
+                return
+            if self.target != "AdjClose":
+                print("MC Prediction designed for AdjClose prices")
+                return
+
+            mc_view.display_mc_forecast(
+                data=self.stock[self.target],
+                n_future=ns_parser.n_days,
+                n_sims=ns_parser.n_sims,
+                use_log=ns_parser.dist == "lognormal",
+                export=ns_parser.export,
+            )
+        except Exception as e:
+            print(e, "\n")
+
 
 def menu(ticker: str, start: datetime, interval: str, stock: pd.DataFrame):
     """Prediction Techniques Menu"""

website/content/stocks/prediction_techniques/mc/_index.md

@@ -0,0 +1,20 @@
+```
+usage: mc [-d N_DAYS] [-n N_SIMS] [--dist {normal,lognormal}] [-h] [--export {png,jpg,pdf,svg}]
+```
+
+Perform Monte Carlo forecasting
+
+```
+optional arguments:
+  -d N_DAYS, --days N_DAYS
+                        Days to predict (default: 30)
+  -n N_SIMS, --num N_SIMS
+                        Number of simulations to perform (default: 100)
+  --dist {normal,lognormal}
+                        Whether to model returns or log returns (default: lognormal)
+  -h, --help            show this help message (default: False)
+  --export {png,jpg,pdf,svg}
+                        Export figure into png, jpg, pdf, svg (default: )
+```
+
+![MC_fig](https://user-images.githubusercontent.com/18151143/138540365-55a253f4-e02f-406d-9398-984d988888d7.png)

website/data/menu/main.yml

@@ -508,6 +508,8 @@ main:
             ref: "/stocks/prediction_techniques/lstm"
           - name: conv1d
             ref: "/stocks/prediction_techniques/conv1d"
+          - name: mc
+            ref: "/stocks/prediction_techniques/mc"
       - name: report
         ref: "/stocks/report"
   - name: forex