@astermind/astermind-elm/dist/ml/TFIDF.d.ts

JavaScript Extreme Learning Machine (ELM) library for browser and Node.js.

export declare class TFIDF {
    termFrequency: Record<string, number>;
    inverseDocFreq: Record<string, number>;
    wordsInDoc: string[];
    processedWords: string[];
    scores: Record<string, number>;
    corpus: string;
    constructor(corpusDocs: string[]);
    static countDocsContainingTerm(corpusDocs: string[], term: string): number;
    static processWords(words: string[]): string[];
    static removeStopWordsAndStem(words: string[]): string[];
    static advancedStem(word: string): string;
    static lemmatize(word: string): string;
    static generateNGrams(tokens: string[], n: number): string[];
    calculateScores(): Record<string, number>;
    extractKeywords(topN: number): Record<string, number>;
    processedWordsIndex(word: string): number;
}
export declare class TFIDFVectorizer {
    vocabulary: string[];
    tfidf: TFIDF;
    docTexts: string[];
    constructor(docs: string[], maxVocabSize?: number);
    /**
     * Returns the dense TFIDF vector for a given document text.
     */
    vectorize(doc: string): number[];
    /**
     * Returns vectors for all original training docs.
     */
    vectorizeAll(): number[][];
    /**
     * Optional L2 normalization utility.
     */
    static l2normalize(vec: number[]): number[];
}