Beginning of a LocationChoiceModel class. (Implicitly MNL.)

urbansim/models/lcm.py

@@ -0,0 +1,71 @@
+import numpy as np
+from patsy import dmatrix
+
+from ..urbanchoice import interaction, mnl
+
+
+class LocationChoiceModel(object):
+    """
+    A location choice model with the ability to store an estimated
+    model and predict new data based on the model.
+
+    Parameters
+    ----------
+    fit_filters : list of str
+        Filters applied before fitting the model.
+    predict_filters : list of str
+        Filters applied before calculating new data points.
+    model_expression : str
+        A patsy model expression. Should contain only a right-hand side.
+    sample_size : int
+        Number of choices to sample for estimating the model.
+    name : optional
+        Optional descriptive name for this model that may be used
+        in output.
+
+    """
+    def __init__(self, fit_filters, predict_filters, model_expression,
+                 sample_size, name=None):
+        self.fit_filters = fit_filters
+        self.predict_filters = predict_filters
+        # LCMs never have a constant
+        self.model_expression = model_expression + ' - 1'
+        self.sample_size = sample_size
+        self.name = name or 'LocationChoiceModel'
+
+    def fit(self, choosers, alternatives, current_choice):
+        """
+        Fit and save model parameters based on given data.
+
+        Parameters
+        ----------
+        choosers : pandas.DataFrame
+            Table describing the agents making choices, e.g. households.
+        alternatives : pandas.DataFrame
+            Table describing the things from which agents are choosing,
+            e.g. buildings.
+        current_choice : pandas.Series
+            A Series describing the `alternatives` currently chosen
+            by the `choosers`. Should have an index matching `choosers`
+            and values matching the index of `alternatives`.
+
+        Returns
+        -------
+        null_ll : float
+            Null Log-liklihood
+        conv_ll : float
+            Log-liklihood at convergence
+        ll_ratio : float
+            Log-liklihood ratio
+
+        """
+        _, merged, chosen = interaction.mnl_interaction_dataset(
+            choosers, alternatives, self.sample_size, current_choice)
+        model_design = dmatrix(self.model_expression, data=merged)
+        fit, results = mnl.mnl_estimate(
+            model_design.as_matrix(), chosen, self.sample_size)
+        self.fit_results = results
+        return fit
+
+    def predict(self, data):
+        pass

urbansim/urbanchoice/interaction.py

@@ -57,6 +57,8 @@ def mnl_simulate(data, coeff, numalts, gpu=GPU, returnprobs=0):
 
 def mnl_interaction_dataset(choosers, alternatives, SAMPLE_SIZE,
                             chosenalts=None):
+    # filter choosers and their current choices if they point to
+    # something that isn't in the alternatives table
     if chosenalts is not None:
         isin = chosenalts.isin(alternatives.index)
         removing = isin.value_counts()[False]
@@ -73,7 +75,7 @@ def mnl_interaction_dataset(choosers, alternatives, SAMPLE_SIZE,
 
     if SAMPLE_SIZE < numalts:
         sample = np.random.choice(
-            alternatives.index.values, SAMPLE_SIZE * choosers.shape[0])
+            alternatives.index.values, SAMPLE_SIZE * numchoosers)
         if chosenalts is not None:
             # replace with chosen alternative
             sample[::SAMPLE_SIZE] = chosenalts
@@ -97,6 +99,7 @@ def mnl_interaction_dataset(choosers, alternatives, SAMPLE_SIZE,
     try:
         alts_sample['join_index'] = np.repeat(choosers.index, SAMPLE_SIZE)
     except:
+        # TODO: log the error here and re-raise the original exception
         raise Exception(
             "ERROR: An exception here means agents and "
             "alternatives aren't merging correctly")
@@ -108,7 +111,7 @@ def mnl_interaction_dataset(choosers, alternatives, SAMPLE_SIZE,
     chosen = np.zeros((numchoosers, SAMPLE_SIZE))
     chosen[:, 0] = 1
 
-    return sample, alts_sample, ('mnl', chosen)
+    return sample, alts_sample, chosen
 
 
 def mnl_choice_from_sample(sample, choices, SAMPLE_SIZE):