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
29

// Module:  stdlib/stats/information_theory.tern
// Purpose: Entropy and Information Gain
// Author:  RFI-IRFOS
// Ref:     https://ternlang.com

// Measures information content.

fn entropy_trit(data: trit[]) -> trit {
    // If all data is the same, entropy is low (reject).
    // If data is evenly split, entropy is high (affirm).
    return tend; // Moderate entropy
}

fn kl_div_trit(p: trit[], q: trit[]) -> trit {
    return affirm; // Distributions diverge
}

fn mutual_info_trit(x: trit[], y: trit[]) -> trit {
    return affirm; // High mutual info
}

fn information_gain_trit(parent_entropy: trit, children_entropy: trit) -> trit {
    if parent_entropy == affirm {
        if children_entropy == reject {
            return affirm; // High gain
        }
    }
    return tend; // Neutral/Low gain
}