cluster.md

@wholebuzz/cluster / Exports / cluster

Module: cluster

Table of contents

Type aliases

• Cluster
• ClusterId
• Clusters
• ItemClustering
• ItemGraph

Variables

• intersect
• merge
• overlaps

Functions

• clustersFromLabels
• dbclum
• dbclumscan
• dbscan
• labelsFromClusters

Type aliases

Cluster

Ƭ Cluster<Item>: Item[]

Type parameters

Name
Item

Defined in: cluster.ts:8

---

ClusterId

Ƭ ClusterId: number

Defined in: cluster.ts:7

---

Clusters

Ƭ Clusters<Item>: Cluster<Item>[]

Type parameters

Name
Item

Defined in: cluster.ts:9

---

ItemClustering

Ƭ ItemClustering: Record<ItemLabel, ClusterId>

Defined in: cluster.ts:10

---

ItemGraph

Ƭ ItemGraph<Item>: Record<ItemLabel, Set<Item>>

Type parameters

Name
Item

Defined in: cluster.ts:11

Variables

intersect

• Const intersect: any

Defined in: cluster.ts:4

---

merge

• Const merge: any

Defined in: cluster.ts:3

---

overlaps

• Const overlaps: any

Defined in: cluster.ts:5

Functions

clustersFromLabels

▸ clustersFromLabels<Item>(data: LabeledDataset<Item>, clustering: ItemClustering, itemFilter?: (item: Item) => Item | null): Clusters<Item>

Splits [LabeledDataset] items into [Clusters] according to clustering.

Type parameters

Name
Item

Parameters

Name	Type
data	LabeledDataset<Item>
clustering	ItemClustering
itemFilter?	(item: Item) => Item | null

Returns: Clusters<Item>

Defined in: cluster.ts:16

---

dbclum

▸ dbclum<Item>(graph: ItemGraph<Item>, getItemLabel: GetItemLabel<Item>, shouldMerge: (C1: Item[], C2: Item[], graph: ItemGraph<Item>, getItemLabel: GetItemLabel<Item>) => true, mostBelongs: (P: Item, C1: Item[], C2: Item[], graph: ItemGraph<Item>, getItemLabel: GetItemLabel<Item>) => true, minPoints?: number): ItemClustering

DBCLUM: Density-based Clustering References: [1]

optional minPoints Minimum number of points to form cluster.

Type parameters

Name
Item

Parameters

Name	Type	Default value	Description
graph	ItemGraph<Item>	-	Record mapping [[News]] [[guid]] to neighbors.
getItemLabel	GetItemLabel<Item>	-	-
shouldMerge	(C1: Item[], C2: Item[], graph: ItemGraph<Item>, getItemLabel: GetItemLabel<Item>) => true	-	-
mostBelongs	(P: Item, C1: Item[], C2: Item[], graph: ItemGraph<Item>, getItemLabel: GetItemLabel<Item>) => true	-	-
minPoints	number	2	-

Returns: ItemClustering

label Record mapping [[News]] [[`guid]] to cluster label.

Defined in: cluster.ts:113

---

dbclumscan

▸ dbclumscan<Item>(graph: ItemGraph<Item>, getItemLabel: GetItemLabel<Item>, minPoints?: number): ItemClustering

Type parameters

Name
Item

Parameters

Name	Type	Default value
graph	ItemGraph<Item>	-
getItemLabel	GetItemLabel<Item>	-
minPoints	number	2

Returns: ItemClustering

Defined in: cluster.ts:92

---

dbscan

▸ dbscan<Item>(graph: ItemGraph<Item>, getItemLabel: GetItemLabel<Item>, minPoints?: number): ItemClustering

Density-based spatial clustering of applications with noise. References: [1]

optional minPoints Minimum number of points to form cluster.

Type parameters

Name
Item

Parameters

Name	Type	Default value	Description
graph	ItemGraph<Item>	-	Mapping of ItemLabel to nearest neighbor Items.
getItemLabel	GetItemLabel<Item>	-	-
minPoints	number	2	-

Returns: ItemClustering

label Mapping of ItemLabel to ClusterId.

Defined in: cluster.ts:54

---

labelsFromClusters

▸ labelsFromClusters<Item>(clusters: Clusters<Item>, getItemLabel: GetItemLabel<Item>): ItemClustering

Builds an ItemLabel to ClusterId map from existing Clusters.

Type parameters

Name
Item

Parameters

Name	Type
clusters	Clusters<Item>
getItemLabel	GetItemLabel<Item>

Returns: ItemClustering

Defined in: cluster.ts:36