svm-techml-smile.classification — net.clojars.alex314159/svm-techml-smile 0.1.1

Time-series SVM classification helpers: label conversion, evaluation metrics, and a generic model spec constructor (clibtrader-style).

A time-series classification model spec is a map with: :dataset — tech.ml.dataset (rows = trading days, columns = OHLCV + derived) :seed — long, for deterministic train/test shuffle :start-idx — skip this many rows at the start (warm-up period for indicators) :last-idx — reserve this many rows at the end for out-of-sample / live prediction :training — fraction of middle rows used for training (rest = test) :id-name — column keyword for row identifiers (e.g. :date) :signal — {:column <kw> :map-fn <fn> :arg-cols [<kw> ...]} target label spec :transform — seq of feature specs applied to build derived columns :features — seq of feature specs (with :lag) used as the feature matrix

evaluate-classifier clj

(evaluate-classifier kernel-machine features-array y-array)

Run a trained Smile SVM on features-array and return predictions + metrics.

kernel-machine — smile.classification.SVM (or compatible .predict implementor) features-array — double[][] y-array — int[] of true labels in {-1, +1}

Returns a map with :predictions (seq), :y-array, :accuracy, :precision, :sensitivity, :fallout, :confusion-matrix.

labels->01 clj

(labels->01 coll)

Convert SVM {-1, +1} integer labels to {0, 1} for Smile classification metrics. Smile's Accuracy/Precision/Sensitivity/Fallout/ConfusionMatrix expect 0-based class indices — passing -1 causes ArrayIndexOutOfBoundsException.

svm-model-spec clj

(svm-model-spec
  dataset
  &
  {:keys [seed start-idx last-idx training id-name]
   :or {seed 42 start-idx 100 last-idx 250 training 0.7 id-name :date}})

Build a base time-series SVM classification spec.

Required args: dataset — tech.ml.dataset (one row per time step)

Optional kwargs (all have defaults): :seed — random seed for deterministic shuffle (default 42) :start-idx — warm-up rows to skip at the start (default 100) :last-idx — rows reserved at the end for live prediction (default 250) :training — train fraction of middle rows (default 0.7) :id-name — row-ID column keyword (default :date)

Callers must assoc :signal, :transform, and :features before passing to the processing pipeline (process-model / split-data / svm-prepare-data in the host project).