Skip to content

fix: build histogram x-axis as intervals (a, b] (#4)#5

Open
luigidcsoares wants to merge 3 commits intomainfrom
fix/4-histogram-bins
Open

fix: build histogram x-axis as intervals (a, b] (#4)#5
luigidcsoares wants to merge 3 commits intomainfrom
fix/4-histogram-bins

Conversation

@luigidcsoares
Copy link
Copy Markdown
Collaborator

@luigidcsoares luigidcsoares commented Apr 17, 2024

Closes #4

This PR changes the way histograms are constructed by (i) using intervals as labels for the keys and (ii) using ceil instead of rounding to compute the bins. By using ceil, all values are transformed into the endpoint of the corresponding interval, so reconstructing the interval becomes trivial.

@luigidcsoares luigidcsoares requested a review from nunesgh April 17, 2024 05:44
@luigidcsoares luigidcsoares self-assigned this Apr 17, 2024
@luigidcsoares luigidcsoares marked this pull request as ready for review April 18, 2024 05:17
@luigidcsoares
Copy link
Copy Markdown
Collaborator Author

The following example illustrates the new format of the histogram. I'm using the synthetic dataset from the single-dataset example. First, to setup and run the analyses (change the path in sys.path.append as appropriate):

import pandas as pd
import sys

sys.path.append("bvm-library")
from bvmlib.bvm import BVM

dataset = pd.DataFrame({
    "id": [i for i in range(1, 11)],
    "age": [25, 25, 25, 25, 25, 49, 49, 49, 49, 60],
    "gender": ["F", "F", "F", "M", "M", "F", "F", "F", "M", "M"],
    "grade": ["A", "A", "C", "B", "B", "C", "C", "E", "D", "D"],
    "disability": [False, True, True, True, False, True, True, False, False, False]
})

bvm = BVM(dataset)
bvm.qids(["age", "gender"])
bvm.sensitive(["grade", "disability"])

results = bvm.assess()

The histogram for re-identification:

results["re_id"].loc[0, "Histogram"]
Output

{'(0, 1]': 0.0,
 '(1, 2]': 0.0,
 '(2, 3]': 0.0,
 '(3, 4]': 0.0,
 '(4, 5]': 0.0,
 '(5, 6]': 0.0,
 '(6, 7]': 0.0,
 '(7, 8]': 0.0,
 '(8, 9]': 0.0,
 '(9, 10]': 0.0,
 '(10, 11]': 0.0,
 '(11, 12]': 0.0,
 '(12, 13]': 0.0,
 '(13, 14]': 0.0,
 '(14, 15]': 0.0,
 '(15, 16]': 0.0,
 '(16, 17]': 0.0,
 '(17, 18]': 0.0,
 '(18, 19]': 0.0,
 '(19, 20]': 0.0,
 '(20, 21]': 0.0,
 '(21, 22]': 0.0,
 '(22, 23]': 0.0,
 '(23, 24]': 0.0,
 '(24, 25]': 0.0,
 '(25, 26]': 0.0,
 '(26, 27]': 0.0,
 '(27, 28]': 0.0,
 '(28, 29]': 0.0,
 '(29, 30]': 0.0,
 '(30, 31]': 0.0,
 '(31, 32]': 0.0,
 '(32, 33]': 0.0,
 '(33, 34]': 0.6,
 '(34, 35]': 0.0,
 '(35, 36]': 0.0,
 '(36, 37]': 0.0,
 '(37, 38]': 0.0,
 '(38, 39]': 0.0,
 '(39, 40]': 0.0,
 '(40, 41]': 0.0,
 '(41, 42]': 0.0,
 '(42, 43]': 0.0,
 '(43, 44]': 0.0,
 '(44, 45]': 0.0,
 '(45, 46]': 0.0,
 '(46, 47]': 0.0,
 '(47, 48]': 0.0,
 '(48, 49]': 0.0,
 '(49, 50]': 0.2,
 '(50, 51]': 0.0,
 '(51, 52]': 0.0,
 '(52, 53]': 0.0,
 '(53, 54]': 0.0,
 '(54, 55]': 0.0,
 '(55, 56]': 0.0,
 '(56, 57]': 0.0,
 '(57, 58]': 0.0,
 '(58, 59]': 0.0,
 '(59, 60]': 0.0,
 '(60, 61]': 0.0,
 '(61, 62]': 0.0,
 '(62, 63]': 0.0,
 '(63, 64]': 0.0,
 '(64, 65]': 0.0,
 '(65, 66]': 0.0,
 '(66, 67]': 0.0,
 '(67, 68]': 0.0,
 '(68, 69]': 0.0,
 '(69, 70]': 0.0,
 '(70, 71]': 0.0,
 '(71, 72]': 0.0,
 '(72, 73]': 0.0,
 '(73, 74]': 0.0,
 '(74, 75]': 0.0,
 '(75, 76]': 0.0,
 '(76, 77]': 0.0,
 '(77, 78]': 0.0,
 '(78, 79]': 0.0,
 '(79, 80]': 0.0,
 '(80, 81]': 0.0,
 '(81, 82]': 0.0,
 '(82, 83]': 0.0,
 '(83, 84]': 0.0,
 '(84, 85]': 0.0,
 '(85, 86]': 0.0,
 '(86, 87]': 0.0,
 '(87, 88]': 0.0,
 '(88, 89]': 0.0,
 '(89, 90]': 0.0,
 '(90, 91]': 0.0,
 '(91, 92]': 0.0,
 '(92, 93]': 0.0,
 '(93, 94]': 0.0,
 '(94, 95]': 0.0,
 '(95, 96]': 0.0,
 '(96, 97]': 0.0,
 '(97, 98]': 0.0,
 '(98, 99]': 0.0,
 '(99, 100]': 0.2}

The histogram for attribute inference when the sensitive attribute is grade:

results["att_inf"].loc[0, "Histogram"]
Output

{'(0, 1]': 0.0,
 '(1, 2]': 0.0,
 '(2, 3]': 0.0,
 '(3, 4]': 0.0,
 '(4, 5]': 0.0,
 '(5, 6]': 0.0,
 '(6, 7]': 0.0,
 '(7, 8]': 0.0,
 '(8, 9]': 0.0,
 '(9, 10]': 0.0,
 '(10, 11]': 0.0,
 '(11, 12]': 0.0,
 '(12, 13]': 0.0,
 '(13, 14]': 0.0,
 '(14, 15]': 0.0,
 '(15, 16]': 0.0,
 '(16, 17]': 0.0,
 '(17, 18]': 0.0,
 '(18, 19]': 0.0,
 '(19, 20]': 0.0,
 '(20, 21]': 0.0,
 '(21, 22]': 0.0,
 '(22, 23]': 0.0,
 '(23, 24]': 0.0,
 '(24, 25]': 0.0,
 '(25, 26]': 0.0,
 '(26, 27]': 0.0,
 '(27, 28]': 0.0,
 '(28, 29]': 0.0,
 '(29, 30]': 0.0,
 '(30, 31]': 0.0,
 '(31, 32]': 0.0,
 '(32, 33]': 0.0,
 '(33, 34]': 0.0,
 '(34, 35]': 0.0,
 '(35, 36]': 0.0,
 '(36, 37]': 0.0,
 '(37, 38]': 0.0,
 '(38, 39]': 0.0,
 '(39, 40]': 0.0,
 '(40, 41]': 0.0,
 '(41, 42]': 0.0,
 '(42, 43]': 0.0,
 '(43, 44]': 0.0,
 '(44, 45]': 0.0,
 '(45, 46]': 0.0,
 '(46, 47]': 0.0,
 '(47, 48]': 0.0,
 '(48, 49]': 0.0,
 '(49, 50]': 0.0,
 '(50, 51]': 0.0,
 '(51, 52]': 0.0,
 '(52, 53]': 0.0,
 '(53, 54]': 0.0,
 '(54, 55]': 0.0,
 '(55, 56]': 0.0,
 '(56, 57]': 0.0,
 '(57, 58]': 0.0,
 '(58, 59]': 0.0,
 '(59, 60]': 0.0,
 '(60, 61]': 0.0,
 '(61, 62]': 0.0,
 '(62, 63]': 0.0,
 '(63, 64]': 0.0,
 '(64, 65]': 0.0,
 '(65, 66]': 0.0,
 '(66, 67]': 0.6,
 '(67, 68]': 0.0,
 '(68, 69]': 0.0,
 '(69, 70]': 0.0,
 '(70, 71]': 0.0,
 '(71, 72]': 0.0,
 '(72, 73]': 0.0,
 '(73, 74]': 0.0,
 '(74, 75]': 0.0,
 '(75, 76]': 0.0,
 '(76, 77]': 0.0,
 '(77, 78]': 0.0,
 '(78, 79]': 0.0,
 '(79, 80]': 0.0,
 '(80, 81]': 0.0,
 '(81, 82]': 0.0,
 '(82, 83]': 0.0,
 '(83, 84]': 0.0,
 '(84, 85]': 0.0,
 '(85, 86]': 0.0,
 '(86, 87]': 0.0,
 '(87, 88]': 0.0,
 '(88, 89]': 0.0,
 '(89, 90]': 0.0,
 '(90, 91]': 0.0,
 '(91, 92]': 0.0,
 '(92, 93]': 0.0,
 '(93, 94]': 0.0,
 '(94, 95]': 0.0,
 '(95, 96]': 0.0,
 '(96, 97]': 0.0,
 '(97, 98]': 0.0,
 '(98, 99]': 0.0,
 '(99, 100]': 0.4}

The histogram for attribute inference when the sensitive attribute is disability:

results["att_inf"].loc[1, "Histogram"]
Output

{'(0, 1]': 0.0,
 '(1, 2]': 0.0,
 '(2, 3]': 0.0,
 '(3, 4]': 0.0,
 '(4, 5]': 0.0,
 '(5, 6]': 0.0,
 '(6, 7]': 0.0,
 '(7, 8]': 0.0,
 '(8, 9]': 0.0,
 '(9, 10]': 0.0,
 '(10, 11]': 0.0,
 '(11, 12]': 0.0,
 '(12, 13]': 0.0,
 '(13, 14]': 0.0,
 '(14, 15]': 0.0,
 '(15, 16]': 0.0,
 '(16, 17]': 0.0,
 '(17, 18]': 0.0,
 '(18, 19]': 0.0,
 '(19, 20]': 0.0,
 '(20, 21]': 0.0,
 '(21, 22]': 0.0,
 '(22, 23]': 0.0,
 '(23, 24]': 0.0,
 '(24, 25]': 0.0,
 '(25, 26]': 0.0,
 '(26, 27]': 0.0,
 '(27, 28]': 0.0,
 '(28, 29]': 0.0,
 '(29, 30]': 0.0,
 '(30, 31]': 0.0,
 '(31, 32]': 0.0,
 '(32, 33]': 0.0,
 '(33, 34]': 0.0,
 '(34, 35]': 0.0,
 '(35, 36]': 0.0,
 '(36, 37]': 0.0,
 '(37, 38]': 0.0,
 '(38, 39]': 0.0,
 '(39, 40]': 0.0,
 '(40, 41]': 0.0,
 '(41, 42]': 0.0,
 '(42, 43]': 0.0,
 '(43, 44]': 0.0,
 '(44, 45]': 0.0,
 '(45, 46]': 0.0,
 '(46, 47]': 0.0,
 '(47, 48]': 0.0,
 '(48, 49]': 0.0,
 '(49, 50]': 0.2,
 '(50, 51]': 0.0,
 '(51, 52]': 0.0,
 '(52, 53]': 0.0,
 '(53, 54]': 0.0,
 '(54, 55]': 0.0,
 '(55, 56]': 0.0,
 '(56, 57]': 0.0,
 '(57, 58]': 0.0,
 '(58, 59]': 0.0,
 '(59, 60]': 0.0,
 '(60, 61]': 0.0,
 '(61, 62]': 0.0,
 '(62, 63]': 0.0,
 '(63, 64]': 0.0,
 '(64, 65]': 0.0,
 '(65, 66]': 0.0,
 '(66, 67]': 0.6,
 '(67, 68]': 0.0,
 '(68, 69]': 0.0,
 '(69, 70]': 0.0,
 '(70, 71]': 0.0,
 '(71, 72]': 0.0,
 '(72, 73]': 0.0,
 '(73, 74]': 0.0,
 '(74, 75]': 0.0,
 '(75, 76]': 0.0,
 '(76, 77]': 0.0,
 '(77, 78]': 0.0,
 '(78, 79]': 0.0,
 '(79, 80]': 0.0,
 '(80, 81]': 0.0,
 '(81, 82]': 0.0,
 '(82, 83]': 0.0,
 '(83, 84]': 0.0,
 '(84, 85]': 0.0,
 '(85, 86]': 0.0,
 '(86, 87]': 0.0,
 '(87, 88]': 0.0,
 '(88, 89]': 0.0,
 '(89, 90]': 0.0,
 '(90, 91]': 0.0,
 '(91, 92]': 0.0,
 '(92, 93]': 0.0,
 '(93, 94]': 0.0,
 '(94, 95]': 0.0,
 '(95, 96]': 0.0,
 '(96, 97]': 0.0,
 '(97, 98]': 0.0,
 '(98, 99]': 0.0,
 '(99, 100]': 0.2}

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Histograms: Fix rounding and improve format

1 participant