mastergopote44 commited on
Commit
4dcaccd
1 Parent(s): 08f4213
Files changed (1) hide show
  1. Long-Term-Care-Aggregated-Data.py +190 -0
Long-Term-Care-Aggregated-Data.py ADDED
@@ -0,0 +1,190 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # -*- coding: utf-8 -*-
2
+ """Long-Term-Care-Aggregated-Data.ipynb
3
+
4
+ Automatically generated by Colaboratory.
5
+
6
+ Original file is located at
7
+ https://colab.research.google.com/drive/14YdgB8b4TtNetbHpTstGH7W4DLa3Yyxq
8
+ """
9
+
10
+ from datasets import GeneratorBasedBuilder, DownloadManager, DatasetInfo, Array3D, BuilderConfig, SplitGenerator, Version
11
+ from datasets.features import Features, Value, Sequence
12
+ import datasets
13
+ import pandas as pd
14
+ import json
15
+ import zipfile
16
+ from PIL import Image
17
+ import numpy as np
18
+ import io
19
+ import csv
20
+ import json
21
+ import os
22
+ from typing import List
23
+ import datasets
24
+ import logging
25
+
26
+ _CITATION = """\
27
+ @misc{long_term_care_aggregated_dataset,
28
+ title = {Long-Term Care Aggregated Dataset},
29
+ author = {Kao, Hsuan-Chen (Justin)},
30
+ year = {2024},
31
+ publisher = {Hugging Face},
32
+ url = {https://github.com/justinkao44/STA663_Project_1},
33
+ }
34
+ """
35
+
36
+ _DESCRIPTION = """\
37
+ The Long-Term Care Aggregated Dataset is a collection of insurance data for 'incidence' and 'termination' categories.
38
+ It is compiled from Long Term Care insurance policies data, providing insights into trends and patterns in insurance claims
39
+ and terminations. This dataset can be used for actuarial analysis, risk assessment, and to inform insurance product development.
40
+ """
41
+
42
+ _HOMEPAGE = "https://github.com/justinkao44/STA663_Project_1"
43
+
44
+ _LICENSE = "Apache-2.0"
45
+
46
+ _URLS = {
47
+ "train_incidence": "https://raw.githubusercontent.com/justinkao44/STA663_Project_1/main/train_filtered_incidence_df.csv",
48
+ "train_termination": "https://raw.githubusercontent.com/justinkao44/STA663_Project_1/main/train_filtered_termination_df.csv",
49
+ "validation_incidence": "https://raw.githubusercontent.com/justinkao44/STA663_Project_1/main/validation_filtered_incidence_df.csv",
50
+ "validation_termination": "https://raw.githubusercontent.com/justinkao44/STA663_Project_1/main/validation_filtered_termination_df.csv",
51
+ }
52
+
53
+
54
+ class LongTermCareAggregatedData(datasets.GeneratorBasedBuilder):
55
+ """Dataset for insurance 'incidence' and 'termination' data."""
56
+
57
+ BUILDER_CONFIGS = [
58
+ datasets.BuilderConfig(name="incidence", version=datasets.Version("1.0.0"), description="This part of the dataset includes incidence features"),
59
+ datasets.BuilderConfig(name="termination", version=datasets.Version("1.0.0"), description="This part of the dataset includes termination features"),
60
+ ]
61
+
62
+ def _info(self):
63
+ if self.config.name == "incidence":
64
+ features = datasets.Features({
65
+ "Group_Indicator": datasets.Value("string"),
66
+ "Gender": datasets.Value("string"),
67
+ "Issue_Age_Bucket": datasets.Value("string"),
68
+ "Incurred_Age_Bucket": datasets.Value("string"),
69
+ "Issue_Year_Bucket": datasets.Value("string"),
70
+ "Policy_Year": datasets.Value("string"),
71
+ "Marital_Status": datasets.Value("string"),
72
+ "Premium_Class": datasets.Value("string"),
73
+ "Underwriting_Type": datasets.Value("string"),
74
+ "Coverage_Type_Bucket": datasets.Value("string"),
75
+ "Tax_Qualification_Status": datasets.Value("string"),
76
+ "Inflation_Rider": datasets.Value("string"),
77
+ "Rate_Increase_Flag": datasets.Value("string"),
78
+ "Restoration_of_Benefits": datasets.Value("string"),
79
+ "NH_Orig_Daily_Ben_Bucket": datasets.Value("string"),
80
+ "ALF_Orig_Daily_Ben_Bucket": datasets.Value("string"),
81
+ "HHC_Orig_Daily_Ben_Bucket": datasets.Value("string"),
82
+ "NH_Ben_Period_Bucket": datasets.Value("string"),
83
+ "ALF_Ben_Period_Bucket": datasets.Value("string"),
84
+ "HHC_Ben_Period_Bucket": datasets.Value("string"),
85
+ "NH_EP_Bucket": datasets.Value("string"),
86
+ "ALF_EP_Bucket": datasets.Value("string"),
87
+ "HHC_EP_Bucket": datasets.Value("string"),
88
+ "Region": datasets.Value("string"),
89
+ "Active_Exposure": datasets.Value("float64"),
90
+ "Total_Exposure": datasets.Value("float64"),
91
+ "Claim_Count": datasets.Value("int32"),
92
+ "Count_NH": datasets.Value("int32"),
93
+ "Count_ALF": datasets.Value("int32"),
94
+ "Count_HHC": datasets.Value("int32"),
95
+ "Count_Unk": datasets.Value("int32"),
96
+ })
97
+ elif self.config.name == "termination":
98
+ features = datasets.Features({
99
+ "Gender": datasets.Value("string"),
100
+ "Incurred_Age_Bucket": datasets.Value("string"),
101
+ "Incurred_Year_Bucket": datasets.Value("string"),
102
+ "Claim_Type": datasets.Value("string"),
103
+ "Region": datasets.Value("string"),
104
+ "Diagnosis_Category": datasets.Value("string"),
105
+ "Claim_Duration": datasets.Value("int64"),
106
+ "Exposure": datasets.Value("int64"),
107
+ "Deaths": datasets.Value("int64"),
108
+ "Recovery": datasets.Value("int64"),
109
+ "Terminations": datasets.Value("int64"),
110
+ "Benefit_Expiry": datasets.Value("int64"),
111
+ "Others_Terminations": datasets.Value("int64"),
112
+ })
113
+ else:
114
+ raise ValueError(f"BuilderConfig name not recognized: {self.config.name}")
115
+
116
+ return datasets.DatasetInfo(
117
+ description=_DESCRIPTION,
118
+ features=features,
119
+ supervised_keys=None,
120
+ homepage="https://www.soa.org/resources/experience-studies/2020/2000-2016-ltc-aggregate-database/",
121
+ citation="Please cite this dataset as: Society of Actuaries (SOA). (2020). Long Term Care Insurance Aggregate Experience Data, 2000-2016."
122
+ )
123
+
124
+ def _split_generators(self, dl_manager):
125
+
126
+ downloaded_files = dl_manager.download(_URLS)
127
+
128
+ return [
129
+ datasets.SplitGenerator(
130
+ name="train_incidence",
131
+ gen_kwargs={
132
+ "data_file": downloaded_files["train_incidence"],
133
+ "split": "incidence",
134
+ },
135
+ ),
136
+ datasets.SplitGenerator(
137
+ name="validation_incidence",
138
+ gen_kwargs={
139
+ "data_file": downloaded_files["validation_incidence"],
140
+ "split": "incidence",
141
+ },
142
+ ),
143
+ datasets.SplitGenerator(
144
+ name="train_termination",
145
+ gen_kwargs={
146
+ "data_file": downloaded_files["train_termination"],
147
+ "split": "termination",
148
+ },
149
+ ),
150
+ datasets.SplitGenerator(
151
+ name="validation_termination",
152
+ gen_kwargs={
153
+ "data_file": downloaded_files["validation_termination"],
154
+ "split": "termination",
155
+ },
156
+ ),
157
+ ]
158
+
159
+ def _generate_examples(self, data_file, split):
160
+ # Read the CSV file for the given configuration
161
+ dataframe = pd.read_csv(data_file)
162
+
163
+ # Determine the feature columns based on the configuration
164
+ feature_columns = self._get_feature_columns_by_config()
165
+
166
+ # Yield examples using the determined feature columns
167
+ for idx, row in dataframe.iterrows():
168
+ feature_dict = {column: row[column] for column in feature_columns if column in dataframe.columns}
169
+ yield idx, feature_dict
170
+
171
+ def _get_feature_columns_by_config(self):
172
+ if self.config.name == "incidence":
173
+ return [
174
+ "Group_Indicator", "Gender", "Issue_Age_Bucket", "Incurred_Age_Bucket",
175
+ "Issue_Year_Bucket", "Policy_Year", "Marital_Status", "Premium_Class",
176
+ "Underwriting_Type", "Coverage_Type_Bucket", "Tax_Qualification_Status",
177
+ "Inflation_Rider", "Rate_Increase_Flag", "Restoration_of_Benefits",
178
+ "NH_Orig_Daily_Ben_Bucket", "ALF_Orig_Daily_Ben_Bucket", "HHC_Orig_Daily_Ben_Bucket",
179
+ "NH_Ben_Period_Bucket", "ALF_Ben_Period_Bucket", "HHC_Ben_Period_Bucket",
180
+ "NH_EP_Bucket", "ALF_EP_Bucket", "HHC_EP_Bucket", "Region",
181
+ "Active_Exposure", "Total_Exposure", "Claim_Count", "Count_NH", "Count_ALF", "Count_HHC", "Count_Unk",
182
+ ]
183
+ elif self.config.name == "termination":
184
+ return [
185
+ "Gender", "Incurred_Age_Bucket", "Incurred_Year_Bucket", "Claim_Type",
186
+ "Region", "Diagnosis_Category", "Claim_Duration", "Exposure", "Deaths",
187
+ "Recovery", "Terminations", "Benefit_Expiry", "Others_Terminations",
188
+ ]
189
+ else:
190
+ raise ValueError(f"Config name not recognized: {self.config.name}")