var-techml-neanderthal

VaR/beta/regression primitives for tech.ml.dataset

• Neanderthal workflows.

Why

Historical-VaR, marginal beta and single-factor regression are standard techniques with nothing portfolio-specific about them once you're handed a return series. This library carries the plain-math pieces out of the app that consumes them: quantile/drawdown/VaR, OLS/conditional-beta/marginal-beta decomposition, return/PnL transforms, and futures panama-roll adjustment. Nothing here knows about a bond, a portfolio, or a database — it takes vectors, matrices and tech.ml.datasets and returns numbers.

Same split as svm-techml-smile: the generic technique lives here, the domain wiring (what data feeds it) stays in the app.

Install

[net.clojars.alex314159/var-techml-neanderthal "0.1.1"]

Namespaces

ns	what's in it
var-techml-neanderthal.core	re-exports everything below — require this one ns if you don't care about the split
var-techml-neanderthal.returns	return/PnL series transforms, day/week/month bucketing, one-year/one-year-weeks/one-year-months
var-techml-neanderthal.regression	single-factor OLS, conditional (up/down) beta, marginal-beta decomposition
var-techml-neanderthal.risk	quantile, max drawdown, historical VaR over a return series
var-techml-neanderthal.futures	panama-roll adjustment across a sequence of futures contracts

Note core aliases via a macro: REPL-driven tooling (Calva, CIDER) resolves the docs, purely static analysis (clj-kondo, offline Cursive index) doesn't — require the defining ns there.

Development

lein test