In [1]:

    

%matplotlib inline

# Imports
import matplotlib.pyplot as plt
import pandas
import seaborn
seaborn.set_style("darkgrid")

# Project imports
from legacy_model import *


    

In [2]:

    

# Get raw data
raw_data = get_raw_scdb_data("../data/input/SCDB_Legacy_01_justiceCentered_Citation.csv")


    

In [3]:

    

# Reset output file timestamp per run
file_timestamp = datetime.datetime.now().strftime("%Y%m%d%H%M%S")

# Reset seed per run
numpy.random.seed(0)

# Setup training time period
min_training_years = 25
baseline_window = 10
term_range = range(raw_data["term"].min() + min_training_years,
                   raw_data["term"].max()+1)

# Setup model
m = None
term_count = 0

# Iterate over all terms
for term in term_range:
    # Setup train and test periods
    train_index = (raw_data.loc[:, "term"] < term).values
    baseline_train_index = ((raw_data.loc[:, "term"] < term) & (raw_data.loc[:, "term"] >= (term-baseline_window))).values
    test_index = (raw_data.loc[:, "term"] == term).values
    
    if test_index.sum() == 0:
        continue
    
    # Fit the "baseline" model
    d = sklearn.dummy.DummyClassifier(strategy="most_frequent")
    d.fit(numpy.zeros_like(raw_data.loc[baseline_train_index, :]), 
          (raw_data.loc[baseline_train_index, "justice_outcome_disposition"]).astype(int))
    
    # Store baseline predictions
    raw_data.loc[test_index, "baseline_predicted"] = d.predict(numpy.zeros_like(raw_data.loc[test_index, :]))


    

In [4]:

    

# Get index and outcomes
evaluation_index = raw_data.loc[:, "term"].isin(term_range)
target_actual = (raw_data.loc[evaluation_index, "justice_outcome_disposition"]).astype(int)
target_baseline = raw_data.loc[evaluation_index, "baseline_predicted"].astype(int)
raw_data.loc[evaluation_index, "baseline_correct"] = (target_actual == target_baseline).astype(int)

# Dummy model
print("Dummy model")
print("="*32)
print(sklearn.metrics.classification_report(target_actual, target_baseline))
print(sklearn.metrics.confusion_matrix(target_actual, target_baseline))
print(sklearn.metrics.accuracy_score(target_actual, target_baseline))
print("="*32)
print("")


    

    




Dummy model
================================
             precision    recall  f1-score   support

         -1       0.33      0.01      0.02     39540
          0       0.52      0.77      0.62    113666
          1       0.57      0.46      0.51     93569

avg / total       0.50      0.53      0.48    246775

[[  495 31931  7114]
 [  618 87102 25946]
 [  398 49888 43283]]
0.530361665485
================================

In [5]:

    

# Plot by term
baseline_correct_ts = raw_data.loc[evaluation_index, :].groupby("term")["baseline_correct"].mean()
baseline_correct_ts.plot()


    

    
Out[5]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f5d53e271d0>






    

In [6]:

    

# Get index and outcomes
evaluation_index = raw_data.loc[:, "term"].isin(term_range)
reverse_target_actual = (raw_data.loc[evaluation_index, "justice_outcome_disposition"] > 0).astype(int)
reverse_target_baseline = (raw_data.loc[evaluation_index, "baseline_predicted"] > 0).astype(int)
raw_data.loc[evaluation_index, "baseline_reverse_correct"] = (reverse_target_actual == reverse_target_baseline).astype(int)

# Dummy model
print("Dummy model - Reverse")
print("="*32)
print(sklearn.metrics.classification_report(reverse_target_actual, reverse_target_baseline))
print(sklearn.metrics.confusion_matrix(reverse_target_actual, reverse_target_baseline))
print(sklearn.metrics.accuracy_score(reverse_target_actual, reverse_target_baseline))
print("="*32)
print("")


    

    




Dummy model - Reverse
================================
             precision    recall  f1-score   support

          0       0.70      0.78      0.74    153206
          1       0.57      0.46      0.51     93569

avg / total       0.65      0.66      0.65    246775

[[120146  33060]
 [ 50286  43283]]
0.662259142944
================================

In [7]:

    

# Plot by term
baseline_reverse_correct_ts = raw_data.loc[evaluation_index, :].groupby("term")["baseline_reverse_correct"].mean()
baseline_reverse_correct_ts.plot()


    

    
Out[7]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f5d53e394e0>






    

In [8]:

    

# Get actual and predicted case outcomes
raw_data.loc[:, "justice_outcome_disposition_reverse"] = (raw_data.loc[evaluation_index, "justice_outcome_disposition"] > 0).astype(int)
raw_data.loc[:, "baseline_predicted_reverse"] = (raw_data.loc[evaluation_index, "baseline_predicted"] > 0).astype(int)
docket_baseline_predicted_reverse = (raw_data.loc[evaluation_index, :].groupby("docketId")["baseline_predicted_reverse"].mean() > 0.5).astype(int)
docket_actual_reverse = (raw_data.loc[evaluation_index, :].groupby("docketId")["case_outcome_disposition"].mean() > 0.5).astype(int)


# Dummy model
print("Dummy model - Reverse")
print("="*32)
print(sklearn.metrics.classification_report(docket_actual_reverse, docket_baseline_predicted_reverse))
print(sklearn.metrics.confusion_matrix(docket_actual_reverse, docket_baseline_predicted_reverse))
print("Accuracy:")
print(sklearn.metrics.accuracy_score(docket_actual_reverse, docket_baseline_predicted_reverse))
print("="*32)
print("")


    

    




Dummy model - Reverse
================================
             precision    recall  f1-score   support

          0       0.69      0.81      0.75     16740
          1       0.63      0.47      0.54     11340

avg / total       0.67      0.67      0.66     28080

[[13586  3154]
 [ 5974  5366]]
Accuracy:
0.674928774929
================================

In [9]:

    

# Create merged docket dataframe
docket_df = pandas.concat([docket_actual_reverse, docket_baseline_predicted_reverse], axis=1)
docket_df.columns = ["outcome_actual", "outcome_predicted"]
docket_df.loc[:, "baseline_reverse_correct"] = (docket_df["outcome_actual"] == docket_df["outcome_predicted"]).astype(int)
docket_df = docket_df.join(raw_data.loc[evaluation_index, ["docketId", "term"]].groupby("docketId")["term"].mean())
docket_df.head()


    

    
Out[9]:






  

    

      

      
outcome_actual
      
outcome_predicted
      
baseline_reverse_correct
      
term
    
    

      
docketId
      

      

      

      

    
  
  

    

      
1816-001-01
      
0
      
0
      
1
      
1816
    
    

      
1816-002-01
      
0
      
0
      
1
      
1816
    
    

      
1816-003-01
      
0
      
0
      
1
      
1816
    
    

      
1816-004-01
      
0
      
0
      
1
      
1816
    
    

      
1816-005-01
      
0
      
0
      
1
      
1816
    
  

In [10]:

    

# Plot by term
baseline_case_reverse_correct_ts = docket_df.groupby("term")["baseline_reverse_correct"].mean()
baseline_case_reverse_correct_ts.plot()


    

    
Out[10]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f5d53eea908>






    

In [11]:

    

# Output model results
raw_data.to_csv("../data/output/raw_docket_justice_baseline_model_most_frequent_10.csv.gz", compression="gzip")