Package com.tnsai.memory.advanced

Class MemoryImportanceScorer

java.lang.Object

com.tnsai.memory.advanced.MemoryImportanceScorer

public final class MemoryImportanceScorer
extends Object

Computes importance scores for memory entries based on a weighted combination of recency, frequency, and relevance.

The formula is:

   importance = w1 * recency + w2 * frequency + w3 * relevance
 

where:

• recency = exponential decay based on time since last access
• frequency = normalized access count
• relevance = externally provided relevance score (e.g., from search)

Weights are normalized to sum to 1.0 at construction time.

Constructor Summary

Constructors

Constructor	Description
MemoryImportanceScorer()	Creates a scorer with equal weights (1/3 each).
MemoryImportanceScorer(double recencyWeight, double frequencyWeight, double relevanceWeight)	Creates a scorer with the given weights.

Method Summary

Modifier and Type	Method	Description
double	frequencyWeight()	Returns the frequency weight.
double	recencyWeight()	Returns the recency weight.
double	relevanceWeight()	Returns the relevance weight.
double	score(MemoryEntry entry)	Scores a memory entry without an external relevance signal (relevance=0).
double	score(MemoryEntry entry, double relevance)	Scores a memory entry's importance.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

MemoryImportanceScorer

public MemoryImportanceScorer(double recencyWeight,
 double frequencyWeight,
 double relevanceWeight)

Creates a scorer with the given weights. Weights are normalized so they sum to 1.0.

Parameters:

recencyWeight - weight for recency component

frequencyWeight - weight for frequency component

relevanceWeight - weight for relevance component

MemoryImportanceScorer

public MemoryImportanceScorer()

Creates a scorer with equal weights (1/3 each).

Method Details

score

public double score(MemoryEntry entry,
 double relevance)

Scores a memory entry's importance.

Parameters:

entry - the memory entry

relevance - an external relevance score in [0, 1] (e.g., from vector search)

Returns:

importance score in [0, 1]

score

public double score(MemoryEntry entry)

Scores a memory entry without an external relevance signal (relevance=0).

Parameters:

entry - the memory entry

Returns:

importance score in [0, 1]

recencyWeight

public double recencyWeight()

Returns the recency weight.

frequencyWeight

public double frequencyWeight()

Returns the frequency weight.

relevanceWeight

public double relevanceWeight()

Returns the relevance weight.