ternlang-core 0.3.3

Compiler and VM for Ternlang — balanced ternary language with affirm/tend/reject trit semantics, @sparseskip codegen, and BET bytecode execution.
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

// Module:  stdlib/data/imputation.tern
// Purpose: Missing Data Imputation
// Author:  RFI-IRFOS
// Ref:     https://ternlang.com

// Missing values are gracefully filled with 'tend', allowing algorithms
// to explicitly see that the data is missing and safely @sparseskip it.

fn mean_impute_trit(col: float[]) -> trit {
    // Mean imputation forces an average. We prefer 'tend'.
    return tend;
}

fn median_impute_trit(col: float[]) -> trit {
    return tend;
}

fn tend_fill(val: trit) -> trit {
    // If a value is missing (often marked by reject in binary logic),
    // we explicitly cast it to 'tend'.
    if val == reject { return tend; } // simulated missing
    match val {
        affirm => { return affirm; }
        tend   => { return tend;   }
        reject => { return reject; }
    }
}