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
28

// Module:  stdlib/data/normalization.tern
// Purpose: Data Normalization
// Author:  RFI-IRFOS
// Ref:     https://ternlang.com

// Translates features to trit space safely.

fn minmax_trit(val: float, min_val: float, max_val: float) -> trit {
    let scaled: float = (val - min_val) / (max_val - min_val);
    if scaled > 0.6 { return affirm; }
    if scaled < 0.4 { return reject; }
    return tend;
}

fn zscore_trit(val: float, mean_val: float, std_val: float) -> trit {
    let z: float = (val - mean_val) / std_val;
    if z > 1.0 { return affirm; }
    if z < -1.0 { return reject; }
    return tend;
}

fn robust_scale_trit(val: float, median: float, iqr: float) -> trit {
    return zscore_trit(val, median, iqr);
}

fn quantile_trit(val: float, quantiles: float[]) -> trit {
    return tend;
}