causal-hub 0.0.4

A library for causal models, inference and discovery.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146

mod table;
pub use table::*;

mod trajectory;
use itertools::Either;
pub use trajectory::*;

use crate::types::{Map, Set};

/// A trait for dataset.
pub trait Dataset {
    /// The type of the values.
    type Values;

    /// The values of the dataset.
    ///
    /// # Returns
    ///
    /// A reference to the values.
    ///
    fn values(&self) -> &Self::Values;

    /// The sample size.
    ///
    /// # Notes
    ///
    /// If the dataset is weighted, this should return the sum of the weights.
    ///
    /// # Returns
    ///
    /// The number of samples in the dataset.
    ///
    fn sample_size(&self) -> f64;

    /// Restrict the dataset to the specified variables.
    ///
    /// # Arguments
    ///
    /// * `x` - Set of variables to select.
    ///
    /// # Returns
    ///
    /// A dataset restricted to the specified variables.
    ///
    fn select(&self, x: &Set<usize>) -> Self;
}

/// An enum representing different methods for handling missing data.
#[non_exhaustive]
#[derive(Clone, Copy, Debug)]
pub enum MissingMethod {
    /// List-wise deletion missing handling method.
    LW,
    /// Pair-wise deletion missing handling method.
    PW,
    /// Inverse probability weighting missing handling method.
    IPW,
    /// Augmented inverse probability weighting missing handling method.
    AIPW,
}

/// A trait for incomplete datasets.
pub trait IncDataset: Dataset + Sized {
    /// The type of the missing data indicator.
    type Missing;
    /// The value of the missing data indicator.
    const MISSING: Self::Missing;

    /// The type of the complete dataset.
    type Complete;
    /// The type of the weighted dataset.
    type Weighted;

    /// Get the missing information.
    ///
    /// # Returns
    ///
    /// A reference to the missing information.
    ///
    fn missing(&self) -> &MissingTable;

    /// Apply a missing data handling method to the dataset.
    ///
    /// # Arguments
    ///
    /// * `m` - The missing data handling method to apply.
    /// * `x` - An optional set of variables to consider for missing data handling.
    /// * `pr` - An optional missing mechanism specification.
    ///
    /// # Returns
    ///
    /// Either a complete or weighted dataset.
    ///
    fn apply_missing_method(
        &self,
        m: &MissingMethod,
        x: Option<&Set<usize>>,
        pr: Option<&Map<usize, Set<usize>>>,
    ) -> Either<Self::Complete, Self::Weighted>;

    /// Perform list-wise (LW) deletion to handle missing data.
    ///
    /// # Returns
    ///
    /// A complete dataset obtained via LW deletion.
    ///
    fn lw_deletion(&self) -> Self::Complete;

    /// Perform pair-wise (PW) deletion to handle missing data for the specified columns.
    ///
    /// # Arguments
    ///
    /// * `x` - A set of column indices for PW deletion.
    ///
    /// # Returns
    ///
    /// A complete dataset restricted to the specified columns via PW deletion.
    ///
    fn pw_deletion(&self, x: &Set<usize>) -> Self::Complete;

    /// Perform inverse probability weighting (IPW) deletion to handle missing data for the specified columns.
    ///
    /// # Arguments
    ///
    /// * `x` - A set of column indices for IPW deletion.
    /// * `pr` - The missing data indicators.
    ///
    /// # Returns
    ///
    /// A weighted dataset restricted to the specified columns via IPW deletion.
    ///
    fn ipw_deletion(&self, x: &Set<usize>, pr: &Map<usize, Set<usize>>) -> Self::Weighted;

    /// Perform augmented inverse probability weighting (AIPW) deletion to handle missing data for the specified columns.
    ///
    /// # Arguments
    ///
    /// * `x` - A set of column indices for AIPW deletion.
    /// * `pr` - The missing data indicators.
    ///
    /// # Returns
    ///
    /// A weighted dataset restricted to the specified columns via AIPW deletion.
    ///
    fn aipw_deletion(&self, x: &Set<usize>, pr: &Map<usize, Set<usize>>) -> Self::Weighted;
}