Close Menu
    Trending
    • OpenAI Hires Instacart CEO to Oversee ChatGPT, Applications
    • How AI Is Transforming Creative Industries: From Art to Music to Writing | by AI With Lil Bro | May, 2025
    • How You’ll Feel Reaching Various Millionaire Milestones ($1-$20M)
    • How to Unlock Your Brand’s Potential Through Engaging Content
    • Uh-Uh, Not Guilty | Towards Data Science
    • How Spark Actually Works: Behind the Curtain of Your First .show() | by B V Sarath Chandra | May, 2025
    • Best Practices for Managing a Virtual Medical Receptionist
    • How I Built a Bulletproof Portfolio (And What Most People Get Wrong)
    Finance StarGate
    • Home
    • Artificial Intelligence
    • AI Technology
    • Data Science
    • Machine Learning
    • Finance
    • Passive Income
    Finance StarGate
    Home»Artificial Intelligence»A Practical Guide to BERTopic for Transformer-Based Topic Modeling
    Artificial Intelligence

    A Practical Guide to BERTopic for Transformer-Based Topic Modeling

    FinanceStarGateBy FinanceStarGateMay 8, 2025No Comments15 Mins Read
    Share Facebook Twitter Pinterest LinkedIn Tumblr Reddit Telegram Email
    Share
    Facebook Twitter LinkedIn Pinterest Email


    has a variety of use circumstances within the pure language processing (NLP) area, resembling doc tagging, survey evaluation, and content material group. It falls below the realm of unsupervised studying method, making it a really cost-effective method that reduces the assets required to gather human-annotated knowledge. We’ll dive deeper into BERTopic, a preferred python library for transformer-based matter modeling, to assist us course of monetary information sooner and reveal how the trending subjects change extra time.
    BERTopic consists of 6 core modules that may be personalized to go well with completely different use circumstances. On this article, we’ll look at, experiment with every module individually and discover how they work collectively coherently to provide the tip outcomes.

    BERTopic: Transformer-Primarily based Matter Modeling (until in any other case famous, all pictures are by the creator)

    At a excessive degree, a typical BERTopic structure consists of:

    • Embeddings: rework textual content into vector representations (i.e. embeddings) that seize semantic which means utilizing sentence-transformer fashions.
    • Dimensionality Discount: scale back the high-dimensional embeddings to a lower-dimensional house whereas preserving essential relationships, together with PCA, UMAP …
    • Clustering: group comparable paperwork collectively based mostly on their embeddings with diminished dimensionality to type distinct subjects, together with HDBSCAN, Okay-Means algorithms …
    • Vectorizers: after matter clusters are fashioned, vectorizers convert textual content into numerical options that can be utilized for matter evaluation, together with rely vectorizer, on-line vectorizer …
    • c-TF-IDF: calculate significance scores for phrases inside and throughout matter clusters to establish key phrases.
    • Illustration Mannequin: leverage semantic similarity between the embedding of candidate key phrases and the embedding of paperwork to search out probably the most consultant matter key phrases, together with KeyBERT, LLM-based methods …

    Challenge Overview

    On this sensible utility, we’ll use Topic Modeling to establish trending subjects in Apple monetary information. Utilizing NewsAPI, we gather each day top-ranked Apple inventory information from Google Search and compile them right into a dataset of 250 paperwork, with every doc containing monetary information for one particular day. Nevertheless, this isn’t the primary focus of this text so be at liberty to switch it with your personal dataset. The target is to exhibit easy methods to rework uncooked textual content paperwork containing high Google search outcomes into significant matter key phrases and refine these key phrases to be extra consultant.


    BERTopic’s 6 Elementary Modules

    1. Embeddings

    embeddings

    BERTopic makes use of sentence transformer fashions as its first constructing block, changing sentences into dense vector representations (i.e. embeddings) that seize semantic meanings. These fashions are based mostly on transformer architectures like BERT and are particularly skilled to provide high-quality sentence embeddings. We then compute the semantic similarity between sentences utilizing cosine distance between the embeddings. Widespread fashions embody:

    • all-MiniLM-L6-v2: light-weight, quick, good basic efficiency
    • BAAI/bge-base-en-v1.5: bigger mannequin with robust semantic understanding therefore offers a lot slower coaching and inference pace.

    There are an enormous vary of pre-trained sentence transformers so that you can select from on the “Sentence Transformer” web site and Huggingface model hub. We are able to use a couple of traces of code to load a sentence transformer mannequin and encode the textual content sequences into excessive dimensional numerical embeddings.

    from sentence_transformers import SentenceTransformer
    
    # Initialize mannequin
    mannequin = SentenceTransformer("all-MiniLM-L6-v2")
    
    # Convert sentences to embeddings
    sentences = ["First sentence", "Second sentence"]
    embeddings = mannequin.encode(sentences)  # Returns numpy array of embeddings

    On this occasion, we enter a set of economic information knowledge from October 2024 to March 2025 into the sentence transformer “bge-base-en-v1.5”. As proven within the outcome under. these textual content paperwork are remodeled into vector embedding with the form of 250 rows and every with 384 dimensions.

    embeddings result

    We are able to then feed this sentence transformer to BERTopic pipeline and maintain all different modules because the default settings.

    from sentence_transformers import SentenceTransformer
    from bertopic import BERTopic
    
    emb_minilm = SentenceTransformer("all-MiniLM-L6-v2")
    topic_model = BERTopic(
        embedding_model=emb_minilm,
    )
    
    topic_model.fit_transform(docs)
    topic_model.get_topic_info()

    As the tip outcome, we get the next matter illustration.

    topic result

    In comparison with the extra highly effective and bigger “bge-base-en-v1.5” mannequin, we get the next outcome which is barely extra significant than the smaller “all-MiniLM-L6-v2” mannequin however nonetheless leaves massive room for enchancment.

    One space for enchancment is decreasing the dimensionality, as a result of sentence transformers usually leads to high-dimensional embeddings. As BERTopic depends on evaluating the spatial proximity between embedding house to type significant clusters, it’s essential to use a dimensionality discount method to make the embeddings much less sparse. Due to this fact, we’re going to introduce varied dimensionality discount methods within the subsequent part.

    2. Dimensionality Discount

    dimensionality reduction

    After changing the monetary information paperwork into embeddings, we face the issue of excessive dimensionality. Since every embedding incorporates 384 dimensions, the vector house turns into too sparse to create significant distance measurement between two vector embeddings. Principal Element Evaluation (PCA) and Uniform Manifold Approximation and Projection (UMAP) are frequent methods to cut back dimensionalities whereas preserving the utmost variance within the knowledge. We’ll take a look at UMAP, BERTopic’s default dimensionality discount method, in additional particulars. It’s a non-linear algorithm adopted from topology evaluation that seeks numerous construction throughout the knowledge. It really works by extending a radius outwards from every knowledge level and connecting factors with its shut neighbors. You may dive extra into the UMAP visualization on this web site “Understanding UMAP“.

    UMAP n_neighbours Experimentation

    An essential UMAP parameter is n_neighbours that controls how UMAP balances native and international construction within the knowledge. Low values of n_neighbors will power UMAP to focus on native construction, whereas massive values will take a look at bigger neighborhoods of every level.
    The diagram under exhibits a number of scatterplots demonstrating the impact of various n_neighbors values, with every plot visualizing the embeddings in an 2-dimensional house after making use of UMAP dimensionality discount.

    With smaller n_neighbors values (e.g. n=2, n=5), the plots present extra tightly coupled micro clusters, indicating a concentrate on native construction. As n_neighbors will increase (in direction of n=100, n=150), the factors type extra cohesive international patterns, demonstrating how bigger neighborhood sizes assist UMAP seize broader relationships within the knowledge.

    UMAP experimentation

    UMAP min_dist Experimentation

    The min_dist parameter in UMAP controls how tightly factors are allowed to be packed collectively within the decrease dimensional illustration. It units the minimal distance between factors within the embedding house. A smaller min_dist permits factors to be packed very carefully collectively whereas a bigger min_dist forces factors to be extra scattered and evenly unfold out. The diagram under exhibits an experimentation on min_dist worth from 0.0001 to 1 when setting the n_neighbors=5. When min_dist is ready to smaller values, UMAP emphasizes on preserving native construction whereas bigger values rework the embeddings right into a round form.

    UMAP experimentation

    We determine to set n_neighbors=5 and min_dist=0.01 based mostly on the hyperparameter tuning outcomes, because it types extra distinct knowledge clusters which might be simpler for the next clustering mannequin to course of.

    import umap
    
    UMAP_N = 5
    UMAP_DIST = 0.01
    umap_model = umap.UMAP(
        n_neighbors=UMAP_N,
        min_dist=UMAP_DIST, 
        random_state=0
    )

    3. Clustering

    clustering

    Following the dimensionality discount module, it’s the method of grouping embeddings with shut proximity into clusters. This course of is key to matter modeling, because it categorizes related textual content paperwork collectively by taking a look at their semantic relationships. BERTopic employs HDBSCAN mannequin by default, which has the benefit in capturing constructions with numerous densities. Moreover, BERTopic supplies the pliability of selecting different clustering fashions based mostly on the character of the dataset, resembling Okay-Means (for spherical, equally-sized clusters) or agglomerative clustering (for hirerarchical clusters).

    HDBSCAN Experimentation

    We’ll discover how two essential parameters, min_cluster_size and min_samples, influence the habits of HDBSCAN mannequin.
    min_cluster_size determines the minimal variety of knowledge factors allowed to type a cluster and clusters not assembly the brink are handled as outliers. When setting min_cluster_size too low, you would possibly get many small, unstable clusters which is perhaps noise. If setting it too excessive, you would possibly merge a number of clusters into one, shedding their distinct traits.

    min_samples calculates the space between a degree and its k-th nearest neighbor, figuring out how strict the cluster formation course of is. The bigger the min_samples worth, the extra conservative the clustering turns into, as clusters will likely be restricted to type in dense areas, classifying sparse factors as noise.

    Condensed Tree is a helpful method to assist us determine acceptable values of those two parameters. Clusters that persist for a wide variety of lambda values (proven because the left vertical axis in a condense tree plot) are thought-about steady and extra significant. We favor the chosen clusters to be each tall (extra steady) and broad (massive cluster dimension). We use condensed_tree_ from HDBSCAN to check min_cluster_size from 3 to 50, then visualize the information factors of their vector house, colour coded by the expected cluster labels. As we progress by way of completely different min_cluster_size, we will establish optimum values that group shut knowledge factors collectively.

    On this experimentation, we chosen min_cluster_size=15 because it generates 4 clusters (highlighted in purple within the condensed tree plot under) with good stability and cluster dimension. Moreover the scatterplot additionally signifies cheap cluster formation based mostly on proximity and density.

    Condensed Tree for HDBSCAN min_cluster_size
    Condensed Bushes for HDBSCAN min_cluster_size Experimentation
    Condensed Tree for HDBSCAN min_samples
    Scatterplots for HDBSCAN min_cluster_size Experimentation

    We then perform an identical train to check min_samples from 1 to 80 and chosen min_samples=5. As you may observe from the visuals, the parameters min_samples and min_cluster_size exert distinct impacts on the clustering course of.

    Condensed Bushes for HDBSCAN min_samples Experimentation
    Scatterplots for HDBSCAN min_samples Experimentation
    import hdbscan
    
    MIN_CLUSTER _SIZE= 15
    MIN_SAMPLES = 5
    clustering_model = hdbscan.HDBSCAN(
        min_cluster_size=MIN_CLUSTER_SIZE,
        metric='euclidean',
        cluster_selection_method='eom',
        min_samples=MIN_SAMPLES,
        random_state=0
    )
    
    topic_model = BERTopic(
        embedding_model=emb_bge,
        umap_model=umap_model,
        hdbscan_model=clustering_model, 
    )
    
    topic_model.fit_transform(docs)
    topic_model.get_topic_info()

    Okay-Means Experimentation

    In comparison with HDBSCAN, utilizing Okay-Means clustering permits us to generate extra granular subjects by specifying the n_cluster parameter, consequently, controlling the variety of subjects generated from the textual content paperwork.

    This picture exhibits a collection of scatter plots demonstrating completely different clustering outcomes when various the variety of clusters (n_cluster) from 3 to 50 utilizing Okay-Means. With n_cluster=3, the information is split into simply three massive teams. As n_cluster will increase (5, 8, 10, and so forth.), the information factors are break up into extra granular groupings. General, it types rounded-shape clusters in comparison with HDBSCAN. We chosen n_cluster=8 the place the clusters are neither too broad (shedding essential distinctions) nor too granular (creating synthetic divisions). Moreover, it’s a correct quantity of subjects for categorizing 250 days of economic information. Nevertheless, be at liberty to regulate the code snippet to your necessities if must establish extra granular or broader subjects.

    Scatterplots for Okay-Means n_cluster Experimentation
    from sklearn.cluster import KMeans
    
    N_CLUSTER = 8
    clustering_model = KMeans(
        n_clusters=N_CLUSTER,
        random_state=0
    )
    
    topic_model = BERTopic(
        embedding_model=emb_bge,
        umap_model=umap_model,
        hdbscan_model=clustering_model, 
    )
    
    topic_model.fit_transform(docs)
    topic_model.get_topic_info()

    Evaluating the subject cluster outcomes of Okay-Means and HDBSCAN reveals that Okay-Means produces extra distinct and significant matter representations. Nevertheless, each strategies nonetheless generate many cease phrases, indicating that subsequent modules are essential to refine the subject representations.

    HDBSCAN Output
    HDBSCAN Output
    K-Means Output
    Okay-Means Output

    4. Vectorizer

    vectorizer

    Earlier modules serve the position of grouping paperwork into semantically comparable clusters, and ranging from this module the primary focus is to fine-tune the subjects by selecting extra consultant and significant key phrases. BERTopic gives varied Vectorizer choices from the fundamental CountVectorizer to extra superior OnlineCountVectorizer which incrementally replace matter representations. For this train, we’ll experiment on CountVectorizer, a textual content processing software that creates a matrix of token counts out of a set of paperwork. Every row within the matrix represents a doc and every column represents a time period from the vocabulary, with the values exhibiting what number of occasions every time period seems in every doc. This matrix illustration permits machine studying algorithms to course of the textual content knowledge mathematically.

    Vectorizer Experimentation

    We’ll undergo a couple of essential parameters of the CountVectorizer and see how they may have an effect on the subject representations.

    • ngram_range specifies what number of phrases to mix collectively into matter phrases. It’s significantly helpful for paperwork consists of quick phrases, which isn’t wanted on this state of affairs.
      instance output if we set ngram_range=(1, 3)
    0                -1_apple nasdaq aapl_apple stock_apple nasdaq_nasdaq aapl   
    1  0_apple warren buffett_apple stock_berkshire hathaway_apple nasdaq aapl   
    2           1_apple nasdaq aapl_nasdaq aapl apple_apple stock_apple nasdaq   
    3              2_apple aapl stock_apple nasdaq aapl_apple stock_aapl inventory   
    4           3_apple nasdaq aapl_cramer apple aapl_apple nasdaq_apple inventory 
    • stop_words determines whether or not cease phrases are faraway from the subjects, which considerably improves matter representations.
    • min_df and max_df determines the frequency thresholds for phrases to be included within the vocabulary. min_df units the minimal variety of paperwork a time period should seem whereas max_df units the utmost doc frequency above which phrases are thought-about too frequent and discarded.

    We discover the impact of including CountVectorizer with max_df=0.8 (i.e. ignore phrases showing in additional than 80% of the paperwork) to each HDBSCAN and Okay-Means fashions from the earlier step.

    from sklearn.feature_extraction.textual content import CountVectorizer
    vectorizer_model = CountVectorizer(
    		max_df=0.8, 
    		stop_words="english"
    )
    
    topic_model = BERTopic(
        embedding_model=emb_bge,
        umap_model=umap_model,
        hdbscan_model=clustering_model, 
        vectorizer_model=vectorizer_model
    )

    Each exhibits enhancements after introducing the CountVectorizer, considerably decreasing key phrases incessantly appeared in all paperwork and never bringing additional values, resembling “appl”, “inventory”, and “apple”.

    HDBSCAN Output with Vectorizer
    HDBSCAN Output with Vectorizer
    K-Means Output with Vectorizer
    Okay-Means Output with Vectorizer

    5. c-TF-IDF

    c-TF-IDF

    Whereas the Vectorizer module focuses on adjusting the subject illustration on the doc degree, c-TF-IDF primarily take a look at the cluster degree to cut back incessantly encountered subjects throughout clusters. That is achieved by changing all paperwork belonging to 1 cluster as a single doc and calculated the key phrase significance based mostly on the standard TF-IDF method.

    c-TF-IDF Experimentation

    • reduce_frequent_words: determines whether or not to down-weight incessantly occurring phrases throughout subjects
    • bm25_weighting: when set to True, makes use of BM25 weighting as an alternative of normal TF-IDF, which may also help higher deal with doc size variations. In smaller datasets, this variant will be extra sturdy to cease phrases.

    We use the next code snippet so as to add c-TF-IDF (with bm25_weighting=True) into our BERTopic pipeline.

    from bertopic.vectorizers import ClassTfidfTransformer
    
    ctfidf_model = ClassTfidfTransformer(bm25_weighting=True)
    topic_model = BERTopic(
        embedding_model=emb_bge,
        umap_model=umap_model,
        hdbscan_model=clustering_model, 
        vectorizer_model=vectorizer_model,
        ctfidf_model=ctfidf_model
    )

    The subject cluster outputs under present that including c-TF-IDF has no main influence to the tip outcomes when CountVectorizer has already been added. That is probably as a result of our CountVectorizer has already set a excessive bar of eliminating phrases showing in additional than 80% on the doc degree. Subsequently, this already reduces overlapping vocabularies on the matter cluster degree, which is what c-TF-IDF is meant to realize.

    HDBSCAN Output with Vectorizer and c-TF-IDF
    Okay-Means Output with Vectorizer and c-TF-IDF

    Nevertheless, If we substitute CountVectorizer with c-TF-IDF, though the outcome under exhibits slight enhancements in comparison with when each should not added, there are too many cease phrases current, making the subject representations much less beneficial. Due to this fact, it seems that for the paperwork we’re coping with on this situation, c-TF-IDF module doesn’t convey additional worth.

    HDBSCAN Output with c-TF-IDF solely
    Okay-Means Output with c-TF-IDF solely

    6. Illustration Mannequin

    The final module is the illustration mannequin which has been noticed having a major influence on tuning the subject representations. As a substitute of utilizing the frequency based mostly method like Vectorizer and c-TF-IDF, it leverages semantic similarity between the embeddings of candidate key phrases and the embeddings of paperwork to search out probably the most consultant matter key phrases. This can lead to extra semantically coherent matter representations and decreasing the variety of synonymically comparable key phrases. BERTopic additionally gives varied customization choices for illustration fashions, together with however not restricted to the next:

    • KeyBERTInspired: make use of KeyBERT method to extract matter phrases based mostly semantic similarity.
    • ZeroShotClassification: benefit from open-source transformers within the Huggingface model hub to assign labels to subjects.
    • MaximalMarginalRelevance: lower synonyms in subjects (e.g. inventory and shares).

    KeyBERTInspired Experimentation

    We discovered that KeyBERTInspired is a really cost-effective method because it considerably improves the tip outcome by including a couple of additional traces of code, with out the necessity of in depth hyperparameter tuning.

    from bertopic.illustration import KeyBERTInspired
    
    representation_model = KeyBERTInspired()
    
    topic_model = BERTopic(gh
        embedding_model=emb_bge,
        umap_model=umap_model,
        hdbscan_model=clustering_model, 
        vectorizer_model=vectorizer_model,
        representation_model=representation_model
    )

    After incorporating the KeyBERT-Impressed illustration mannequin, we now observe that each fashions generate noticeably extra coherent and beneficial themes.

    HDBSCAN Output with KeyBERTInspired
    HDBSCAN Output with KeyBERTInspired
    K-Means Output with KeyBERTInspired
    Okay-Means Output with KeyBERTInspired

    Take-Residence Message

    This text explores BERTopic method and implementation for matter modeling, detailing its six key modules with sensible examples utilizing Apple inventory market information knowledge to exhibit every part’s influence on the standard of matter representations.

    • Embeddings: use transformer-based embedding fashions to transform paperwork into numerical representations that seize semantic which means and contextual relationships in textual content.
    • Dimensionality Discount: make use of UMAP or different dimensionality discount methods to cut back high-dimensional embeddings whereas preserving each native and international construction of the information
    • Clustering: examine HDBSCAN (density-based) and Okay-Means (centroid-based) clustering algorithm to group comparable paperwork into coherent subjects
    • Vectorizers: use Rely Vectorizer to create document-term matrices and refine subjects based mostly on statistical method.
    • c-TF-IDF: replace matter representations by analyzing time period frequency at cluster degree (matter class) and scale back frequent phrases throughout completely different subjects.
    • Illustration Mannequin: refine matter key phrases utilizing semantic similarity, providing choices like KeyBERTInspired and MaximalMarginalRelevance for higher matter descriptions



    Source link

    Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
    Previous ArticleHow to build a better AI benchmark
    Next Article How I Built a Bulletproof Portfolio (And What Most People Get Wrong)
    FinanceStarGate

    Related Posts

    Artificial Intelligence

    Uh-Uh, Not Guilty | Towards Data Science

    May 8, 2025
    Artificial Intelligence

    Generating Data Dictionary for Excel Files Using OpenPyxl and AI Agents

    May 8, 2025
    Artificial Intelligence

    Real-Time Interactive Sentiment Analysis in Python

    May 8, 2025
    Add A Comment
    Leave A Reply Cancel Reply

    Top Posts

    Generative AI and Civic Institutions

    March 4, 2025

    SoftBank to Spend $3B Annually on OpenAI Solutions

    February 3, 2025

    Think Out Loud: Rethinking AI with Reinforcement Learning | by Hammad Abbasi | Feb, 2025

    February 1, 2025

    How AI-Powered Healthcare Solutions Transforming Telemedicine?

    February 13, 2025

    xnwochsjhd – mibuv nicecjg – Medium

    March 21, 2025
    Categories
    • AI Technology
    • Artificial Intelligence
    • Data Science
    • Finance
    • Machine Learning
    • Passive Income
    Most Popular

    OSEMN framework overview – 桜満 集

    February 13, 2025

    Optimizing Linear Programming: CVXOPT vs. Geometric Approach — A Hands-on Guide | by Trumphan | Feb, 2025

    February 25, 2025

    Least Squares: Where Convenience Meets Optimality

    March 25, 2025
    Our Picks

    Why Entrepreneurs Are Becoming the New Politicians

    May 6, 2025

    Sacrificing The Stock Market For The Good Of Your Loving Home

    April 7, 2025

    A Beginner’s Guide to Neutral Networks | by Konna Giann | Apr, 2025

    April 3, 2025
    Categories
    • AI Technology
    • Artificial Intelligence
    • Data Science
    • Finance
    • Machine Learning
    • Passive Income
    • Privacy Policy
    • Disclaimer
    • Terms and Conditions
    • About us
    • Contact us
    Copyright © 2025 Financestargate.com All Rights Reserved.

    Type above and press Enter to search. Press Esc to cancel.