{"id":3364,"date":"2023-12-17T20:37:54","date_gmt":"2023-12-17T11:37:54","guid":{"rendered":"https:\/\/saraheee.com\/?p=3364"},"modified":"2023-12-17T20:37:56","modified_gmt":"2023-12-17T11:37:56","slug":"dp3-database-anonymization-randomized-response","status":"publish","type":"post","link":"https:\/\/saraheee.com\/ko\/2023\/12\/dp3-database-anonymization-randomized-response\/","title":{"rendered":"[DP#3] Database anonymization \u2013\u00a0randomized response"},"content":{"rendered":"

Randomized Response Explained<\/h2>\n\n\n\n

Did you ever cheat on your spouse?
Are you using any illegal drugs?
Have you committed any crimes in your past?<\/p>\n\n\n\n

Randomized Response<\/h4>\n\n\n\n

S.L. Warner, 1965
– I never said that!<\/p>\n\n\n\n

How it works<\/h4>\n\n\n\n

scenario respondents are asked a sensitive yes or no question a binary question
– Have you ever used illegal drugs?<\/mark><\/p>\n\n\n\n

instead of answering directly respondents are ask to flip a coin in private
if the coin comes up heads they answer truthfully
if it comes up Tails respondents flip the coin again and answer yes if it lands hats and no if it lands tails
in reality you would flip the coin twice in both scenarios<\/mark><\/p>\n\n\n\n

Data Collection<\/h4>\n\n\n\n

Real answer = “Yes”
3\/4 = 75%<\/p>\n\n\n\n

Advantages<\/h4>\n\n\n\n

Increased Truthfulness: \uc751\ub2f5\uc790\uac00 \ubbfc\uac10\ud55c \uc9c8\ubb38\uc5d0 \ub300\ud574 \ubcf4\ub2e4 \uc9c4\uc2e4\ud55c \ub2f5\ubcc0\uc744 \uc81c\uacf5\ud558\ub3c4\ub85d \uc7a5\ub824\ud568(\uc2e4\uc81c \uc751\ub2f5\uc740 \ubb34\uc791\uc704\uc131\uc5d0 \uc758\ud574 \uac00\ub824\uc9c0\ubbc0\ub85c)
Privacy Protection: \uc124\ubb38\uc870\uc0ac \ub370\uc774\ud130\uc5d0 \uc811\uadfc\ud558\ub354\ub77c\ub3c4 \uc751\ub2f5\uc790\uc758 \uac1c\uc778\uc815\ubcf4 \ubcf4\ud638\ub97c \ubcf4\uc7a5\ud568
Statistical Analysis: \ud1b5\uacc4 \ubd84\uc11d\ud558\uc5ec \uc778\uad6c \uc9d1\ub2e8 \ub0b4\uc758 \ubbfc\uac10\ud55c \ud589\ub3d9\uc774\ub098 \ud2b9\uc131\uc758 \uc720\ubcd1\ub960 \ucd94\uc815<\/p>\n\n\n\n

Challenges<\/h4>\n\n\n\n

Randomization Device Choice
Sample Size
Assumption of Randomness
Ethical Concerns<\/p>\n\n\n\n

Applications<\/h4>\n\n\n\n

Substance abuse
Criminal behavior
Sexual preferences
Local Differential Privacy
RAPPOR<\/p>\n\n\n\n

Conclusion<\/h4>\n\n\n\n

Collecting sensitive information
Preserving privacy
Enhance the quality<\/p>\n\n\n\n

km-anonymity<\/h2>\n\n\n\n
Name<\/td>Age<\/td>Sex<\/td>ZIP<\/td>Disease<\/td><\/tr>
Alice<\/td>28<\/td>F<\/td>23467<\/td>Cancer<\/td><\/tr>
Bob<\/td>17<\/td>M<\/td>12345<\/td>Heart disease<\/td><\/tr>
Charly<\/td>34<\/td>M<\/td>65490<\/td>Flu<\/td><\/tr>
Dave<\/td>41<\/td>M<\/td>84933<\/td>Bronchitis<\/td><\/tr><\/tbody><\/table>
Identifiers: Name – removed
Quasi Identifiers: Age, Sex, ZIP
Sensitive Attribute: Disease<\/figcaption><\/figure>\n\n\n\n
Purchases<\/td>Purchases(anonymity)<\/td><\/tr>
Cherries, Apples, Grapes, Tea, Coffee, Diapers<\/td>Cherries, Apples, Grapes, Tea, Coffee, Diapers<\/td><\/tr>
Flour, Coffee, Bread, Butter<\/td>Cherries, Apples, Grapes, Bread, Butter<\/td><\/tr>
Milk, Yoghurt, Brie, Stilton<\/td>Milk, Yoghurt, Brie, Stilton<\/td><\/tr>
Jam, Gouda, Ham, Bacon, Pepperoni<\/td>Cherries, Apples, Grapes, Gouda, Ham, Bacon, Pepperoni<\/td><\/tr>
Tomatoes, Brokkoli, Potatoes, Pasta, Rice<\/td>Cherries, Apples, Grapes, Brokkoli, Potatoes, Pasta, Rice<\/td><\/tr>
Soft Drink, Beer, Fish, Meat, Eggs<\/td>Soft Drink, Beer, Fish, Meat, Eggs<\/td><\/tr><\/tbody><\/table>
\\(k^3\\)-anonymity: same m items(m=3, Cherries, Apples, Grapes)
\\(4^3\\)-anonymity<\/figcaption><\/figure>\n\n\n\n

Generalization<\/h4>\n\n\n
\n
\"\"<\/figure><\/div>\n\n\n

I have cherries, apples and grapes which can be generalized to Fruit<\/p>\n\n\n\n

Purchases(anonymity)<\/td>Purchases(generalization)<\/td><\/tr>
Cherries, Apples, Grapes, Tea, Coffee, Diapers<\/td>Fruits, Beverages, Household Items<\/td><\/tr>
Cherries, Apples, Grapes, Bread, Butter<\/td>Fruits, Bread, Dairy<\/td><\/tr>
Milk, Yoghurt, Brie, Stilton<\/td>Dairy<\/td><\/tr>
Cherries, Apples, Grapes, Gouda, Ham, Bacon, Pepperoni<\/td>Fruits, Dairy, Sausages<\/td><\/tr>
Cherries, Apples, Grapes, Brokkoli, Potatoes, Pasta, Rice<\/td>Fruits, Vegetables, Pasta<\/td><\/tr>
Soft Drink, Beer, Fish, Meat, Eggs<\/td>Beverages, Meats<\/td><\/tr><\/tbody><\/table>
trade-off between Accuracy and Privacy<\/figcaption><\/figure>\n\n\n\n

Step-By-Step Guide on Implementing Multi-Dimensional Mondrian to Achieve k-anonymity<\/h2>\n\n\n\n

python mondrian_youtube.py<\/p>\n\n\n\n

import pandas as pd\n\ndata = pd.read_csv('data.csv')\nk = 3\n\nqis = ['Age', 'ZIP', 'Gender']\n\ndef summarized(partition, dim):\n    for qi in qis:\n        partition = partition.sort_values(by=qi)\n\n        if(partition[qi].iloc[0] != partition[qi].iloc[-1]):\n            s = f\"[{partition[qi].iloc[0]} - {partition[qi].iloc[-1]}]\"\n            partition[qi] = [s]*partition[qi].size\n    return partition\n\ndef anonymize(partitions, ranks):\n    dim = ranks[0][0]\n\n    partition = partition.sort_values(by=dim)\n    si = partition[dim].count()\n    mid = si\/\/2\n\n    lhs = partition[:mid]\n    rhs = partition[mid:]\n\n    if(len(lhs)) >= k and len(rhs) >= k):\n        return pd.concat([anonymize(lhs, ranks), anonymize(rhs, ranks)])\n    return summarized(partition, dim)\n\ndef mondrian(partitions):\n    ranks = {}\n\n    for qi in qis:\n        ranks[qi] = len(set(partition[qi]))\n\n    # sort ranks\n    ranks = sorted(ranks.items(), key=lambda t: t[1], reverse=True)\n    print(rank)\n\n    return anonymize(partition, ranks)\n\nresult = mondrian(data)\n\nprint(result)\nresult.to_csv('anon_youtube.csv',  index=False)<\/code><\/pre>\n\n\n\n
<\/p>\n\n\n\n
References<\/h3>\n\n\n\n