com.yetanalytics.datasim.timeseries — com.yetanalytics/datasim 0.1.12

arma-seq clj

(arma-seq {:keys [seed] :as arma-model})

(arma-seq {:keys [std phi theta c] :as arma-model :or {phi [] theta []}}
          prev-value
          prev-epsilon
          rng)

Given arma params, return an infinite lazy seq of values

arma-seq-const clj

(arma-seq-const {:keys [std phi theta c seed rng value epsilon]
                 :as arma-model
                 :or {phi [] theta [] value 0.0 epsilon 0.0}}
                n)

Find the value of a stochastic sequence after n runs with the given model.

constant-seq clj

(constant-seq constant)

Return an infinite sequence of the given constant value

cycle-seq clj

(cycle-seq xs & {:keys [length offset] :or {offset 0 length 0}})

Given a sequence, length and offset, return a seq that cycles forever over a portion of the seq.

interval-seq clj

(interval-seq xs)

Return a seq representing the intervals of the input seq

invert-seq clj

(invert-seq xs)

flip vals in a seq from positive to negative or visa versa

op-seq clj

(op-seq op seqs)

Perform actions on one or more seqs

overlap-seq clj

(overlap-seq a
             b
             &
             {:keys [comp-fn extra-stats] :or {comp-fn > extra-stats false}})

NOT USED, but instructive... Given two seqs a and b, for each period where a > b return the T (index) and length of the overlap.

rand-seq clj

(rand-seq &
          {:keys [seed rng val-type gauss-mean gauss-sd]
           :or {val-type :long gauss-mean 0.0 gauss-sd 1.0}})

scale-seq clj

(scale-seq xs scale)

Given a seq and a scale, change the number of events to fit the scale

smooth-seq clj

(smooth-seq xs & {:keys [n] :or {n 2}})

sum-seq clj

(sum-seq & seqs)

Add together the values of any number of seqs

take-sample clj

(take-sample xs sample-millis & {:keys [from] :or {from :millis}})

Take a sample of sample-millis from a time series. :from denotes the period of xs

time-seqs clj

(time-seqs &
           {:keys [t-zero sample-n zone] :or {t-zero 0 zone (t/zone-id "UTC")}})

Given a t-zero (simulation start), an upper bound of sample-n milliseconds and an optional local timezone, return a map of useful lazy time sequences.