model

package

v0.0.0-...-762ccde Latest Latest Go to latest Published: Mar 24, 2018 License: MIT Imports: 7 Imported by: 0

Details

Valid go.mod file

The Go module system was introduced in Go 1.11 and is the official dependency management solution for Go.
Redistributable license

Redistributable licenses place minimal restrictions on how software can be used, modified, and redistributed.
Tagged version

Modules with tagged versions give importers more predictable builds.
Stable version

When a project reaches major version v1 it is considered stable.
Learn more about best practices

Repository

github.com/bobonovski/gotm

Links

Open Source Insights

Documentation ¶

Index ¶

func Register(modelType string, m ModelCtor)
type LDA
type Model
- func NewLDA(topicNum uint32, alpha float32, beta float32) Model
- func NewSparseLDA(topicNum uint32, alpha float32, beta float32) Model
type ModelCtor
- func GetModel(modelType string) (ModelCtor, error)
type SparseLDA

Constants ¶

This section is empty.

Variables ¶

This section is empty.

Functions ¶

func Register ¶

func Register(modelType string, m ModelCtor)

new LDA sampler should register itself using this function

Types ¶

type LDA ¶

type LDA struct {
	Alpha    float32 // document topic mixture hyperparameter
	Beta     float32 // topic word mixture hyperparameter
	TopicNum uint32

	Data *corpus.Corpus // for convenience

	Wt  *sstable.Uint32Matrix      // word-topic count table
	Dt  *sstable.Uint32Matrix      // doc-topic count table
	Wts *sstable.Uint32Matrix      // word-topic-sum count table
	Dwt map[sstable.DocWord]uint32 // doc-word-topic map
}

func (*LDA) Infer ¶

func (this *LDA) Infer(dat *corpus.Corpus, iter int)

infer topics on new documents

func (*LDA) Init ¶

func (this *LDA) Init()

func (*LDA) Likelihood ¶

func (this *LDA) Likelihood() float64

compute the joint likelihood of corpus

func (*LDA) LoadWordTopic ¶

func (this *LDA) LoadWordTopic(fn string) error

deserialize word-topic matrix

func (*LDA) Phi ¶

func (this *LDA) Phi() *sstable.Float32Matrix

compute the posterior point estimation of word-topic mixture beta (Dirichlet prior) + data -> phi

func (*LDA) ResampleTopics ¶

func (this *LDA) ResampleTopics(iter int)

func (*LDA) SavePhi ¶

func (this *LDA) SavePhi(fn string) error

serialize word-topic distribution

func (*LDA) SaveTheta ¶

func (this *LDA) SaveTheta(fn string) error

serialize document-topic distribution

func (*LDA) SaveWordTopic ¶

func (this *LDA) SaveWordTopic(fn string) error

serialize word-topic matrix

func (*LDA) Theta ¶

func (this *LDA) Theta() *sstable.Float32Matrix

compute the posterior point estimation of document-topic mixture alpha (Dirichlet prior) + data -> theta

func (*LDA) Train ¶

func (this *LDA) Train(dat *corpus.Corpus, iter int)

type Model ¶

type Model interface {
	// train model for iter iteration
	Train(dat *corpus.Corpus, iter int)
	// do inference for new doc for iter iteration
	Infer(dat *corpus.Corpus, iter int)
	// get doc-topic distribution
	Phi() *sstable.Float32Matrix
	// get word-topic distribution
	Theta() *sstable.Float32Matrix
	// serialize posterior document topic distribution
	SaveTheta(fn string) error
	// serialize posterior word topic distribution
	SavePhi(fn string) error
	// serialize word topic count table
	SaveWordTopic(fn string) error
	// deserialize word topic count table
	LoadWordTopic(fn string) error
}

the common interface new LDA samplers should follow

func NewLDA ¶

func NewLDA(topicNum uint32, alpha float32, beta float32) Model

New creates a LDA instance with collapsed gibbs sampler

func NewSparseLDA ¶

func NewSparseLDA(topicNum uint32, alpha float32, beta float32) Model

NewSparseLDA creates a sparse lda instance with time and memory efficient gibbs sampler

type ModelCtor ¶

type ModelCtor func(topicNum uint32, alpha float32, beta float32) Model

func GetModel ¶

func GetModel(modelType string) (ModelCtor, error)

type SparseLDA ¶

type SparseLDA struct {
	*LDA
	Wtm *sstable.SortedMap
}

func (*SparseLDA) Infer ¶

func (this *SparseLDA) Infer(dat *corpus.Corpus, iter int)

infer topics on new documents

func (*SparseLDA) Likelihood ¶

func (this *SparseLDA) Likelihood() float64

compute the joint likelihood of corpus

func (*SparseLDA) LoadWordTopic ¶

func (this *SparseLDA) LoadWordTopic(fn string) error

deserialize word-topic matrix

func (*SparseLDA) Phi ¶

func (this *SparseLDA) Phi() *sstable.Float32Matrix

compute the posterior point estimation of word-topic mixture beta (Dirichlet prior) + data -> phi

func (*SparseLDA) ResampleTopics ¶

func (this *SparseLDA) ResampleTopics(iter int)

func (*SparseLDA) SavePhi ¶

func (this *SparseLDA) SavePhi(fn string) error

serialize word-topic distribution

func (*SparseLDA) SaveWordTopic ¶

func (this *SparseLDA) SaveWordTopic(fn string) error

serialize word-topic matrix

func (*SparseLDA) Train ¶

func (this *SparseLDA) Train(dat *corpus.Corpus, iter int)

Source Files ¶

View all Source files

?	: This menu
/	: Search site
f or F	: Jump to
y or Y	: Canonical URL