__
   ____ ___  / /____ _
  / __ `__ \/ // __ `/
 / / / / / / // /_/ /
/_/ /_/ /_/_(_)__, /
                /_/

Machine Learning for kdb+/q

Examples

Reinforcement Learning

• Q-learning applied to trading

DBSCAN

Find clusters of points based on density

q)t:{3?1.} each til 100
q)dbscan[t;3;.2]
-1 0 0 1 0 -1 0 -1 1 1 2 0 3 0 -1 1 4 1 1 1 2 1 0 -1 2 1 2 1 1 4 1 4 1 1 1 0 ..

k nearest neighbors

Find the n closest points to a target

q)t:(`a`b`c!) each {3?100} each til 1000000
q)knn[t;`a`b`c!1 2 3;3]
a b c dist
----------
1 2 3 0
1 2 4 1
0 2 3 1

k-means clustering

Find n centroids in a dataset partitioned by k-means

q)iris:flip `sl`sw`pl`pw`class!("FFFFS";",") 0: `:iris.csv
q)kmeans[delete class from iris;3]
sl       sw       pl       pw
-----------------------------------
6.85     3.073684 5.742105 2.071053
5.901613 2.748387 4.393548 1.433871
5.006    3.418    1.464    0.244

Iris Dataset available here: https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data

Decision Tree (ID3)

q)id3[weatherdata[]]
n_                v_            class
-------------------------------------
,`outlook         ,`overcast    yes
`outlook`wind     `rain`weak    yes
`outlook`wind     `rain`strong  no
`outlook`humidity `sunny`high   no
`outlook`humidity `sunny`normal yes