Crate dtrees_rs 

Decision trees over binary features.

• DL8.5 learns optimal trees by dynamic programming with branch-and-bound and caching (Aglin, Nijssen and Schaus, AAAI 2020). With search rules it becomes anytime: limited discrepancy search (LDS-DL8.5, ECML PKDD 2022), Top-k, and the general CA-DL8.5 framework (IDA 2026). See algorithms::optimal::rules.
• LGDT grows a tree greedily, choosing each test with a depth-2 lookahead (IDA 2024).

Data comes as a Cover, read from a text file by DataReader: one instance per line, the label first, then the 0/1 features.

use dtrees_rs::algorithms::greedy::factories::with_error_minimizer;
use dtrees_rs::algorithms::TreeSearchAlgorithm;
use dtrees_rs::reader::data_reader::DataReader;
use std::path::Path;

let mut cover = DataReader::default().read_file(Path::new("data.txt"))?;
let mut lgdt = with_error_minimizer().max_depth(4).min_support(5).build()?;
lgdt.fit(&mut cover)?;
println!("{}", lgdt.tree());

Modules

algorithms

The tree learners: optimal (optimal: DL8.5 and depth-2 solvers) and greedy with lookahead (greedy: LGDT).

bitsets

Fixed-capacity bitsets.

caching

The cache of subproblems used by DL8.5.

cover

The instances reaching the current node of the search.

globals

Item encoding and small numeric helpers.

parsers

Command line parsing for the examples and the dtrees CLI.

reader

tree

Binary decision trees stored as an arena of nodes.