Our main goal with MonkeyLearn is to make text analysis simple.
We want to enable developers, marketers, salespeople, customer support representatives and entrepreneurs without any knowledge or experience in natural language processing or machine learning to take advantage of this technology.
MonkeyLearn does this through offering models that you can use to analyze your text. Models are very powerful, they can be used for categorizing text or extracting pieces of information, and they can be combined for complex processes. Plus, they improve in accuracy over time.
With these things in mind, users have two basic ways to use MonkeyLearn:
- Public Models: pre-trained and ready to use models for general tasks, created by the MonkeyLearn team. These models do not require any more training, are great at top-level analysis, but they are not as reliable as a Custom Model for more specific situations.
- Custom Models: created by users for their specific needs and based on their own context and criteria. Custom models take into account a client's unique conditions and processes, and can be very accurate with proper training. There are two kinds of models you can make, custom classifiers and custom extractors.
What's the difference between classifiers and extractors? See the explanation here.
By using public models, users can have an easy and fast way to incorporate text analysis capabilities to their apps or workflows. Public models are models that have been already trained with training data and resolve particular text analysis tasks. They are ready to be integrated via a web API, Zapier or our Google Sheets integration.
These models are specialists in common tasks. Click on the link to test the model out.
General Use Classifiers
- NPS SaaS Product Classifier: Classify NPS responses for SaaS products into tags like Ease of Use, Features, Pricing, and Support.
- Urgency Detection: Categorizes incoming pieces of text as urgent or not urgent based on if there is a request for immediate attention, such as "right away, as soon as possible, etc." This will not detect the issue itself (you can build a custom model for that).
- Emotions Classifier: Emotions include anger, boredom, empty, enthusiasm, fun, happiness, hate, love, neutral, relief, sadness, surprise and worry.
- Outbound Sales Response Classifier: Classifies outbound sales email responses based on their subject and body.
- E-commerce Customer Support Ticket Classifier: Classify customer service tickets for e-commece sites into categories like Fraud, Missing Item, Shipping Problem, etc.
- Hotel Aspect: Module for detecting aspects and topics from hotel reviews.
- Retail Classifier: Classifies retail products based on their descriptions.
- Language Classifier: Detect language in text. New languages were added for a total of 49 different languages arranged in language families.
- Topic Classifier: Classifies text by topic.
- Business classifier: Classifies professional profiles, companies or jobs by industry.
- Roles Industry Classifier: Classifies job titles into categories corresponding to industries like "healthcare", "real estate", "software", etc.
- Role Seniority Classifier: Classifies job titles into categories corresponding to seniorities such as "executive-director", "manager", "founder-owner", or "other".
- Role Position Classifier: Classifies job titles into categories corresponding to positions within a company like "marketing", "customer_service", "engineering", etc.
- News Classifier: Classifies news by topics, such as Sports, Politics, Business, and more.
- Startup News - Events Classifier: Given a startup news article, will output what event the piece of news is describing.
- Hacker News Classifier: HackerNews post categorizer. We used data from different subreddits to train the machine learning algorithm.
- Events Classifier: Classify events according to their description
- Sentiment Analysis (English): This is a generic sentiment analysis classifier for texts in English. It works great for any kind of texts. If you are not sure of which sentiment analysis classifier to use (more below), use this one.
- Tweet Sentiment: Classifying tweets in English according to their sentiment: positive, neutral or negative.
- Product Sentiment: Classify product reviews and opinions in English as positive or negative according to their sentiment.
- Hotel Sentiment: Positive/negative sentiment analysis trained over TripAdvisor hotel reviews.
- Restaurant Sentiment: This sentiment analysis classifier was trained with data from different restaurant review sites.
- Movies Sentiment: This sentiment analysis classifier was trained with data from different movie review sites.
- Airline Sentiment: Sentiment analysis for tweets about airline comments.
General Use Extractors
- Keyword Extractor: Extract keywords from text in English. Keywords can be compounded by one or more words and are defined as the important topics in your content and can be used to index data, generate tag clouds or for searching.
- Boilerplate Extractor: Extract relevant text from HTML. This algorithm can be used to detect and remove the surplus "clutter" (boilerplate, templates) around the main textual content of a web page.
- Email Cleaner & Last Reply Extractor: Extract the last reply from an email thread. A cleaning module that removes signatures, confidentiality clauses and other replies from email threads. It relies on statistical algorithms and natural language processing technology to analyze your emails and generate a cleaned version that captures the actual message. The targeted language is English.
- Summary Extractor: Summarize text in English. Given a text, the output will be a shorter version of it that maintains its meaning. This summarization module employs statistical algorithms and natural language processing technology to analyze your content and generate a summary that preserves the gist of the original. No new sentences are generated; every sentence of the summary is present in the original text.
- Opinion Unit Extractor: Extracts opinion units from a given text. Useful to separate paragraphs or sentences into smaller pieces of data.
- Date and Time Extractor: Extracts dates and times from text, and outputs them in ISO format. If a date contains a time, they will be extracted together. When any element of the date is missing (such as the year), the current date is assumed. This base date can be specified as well.
- Email Extractor: Extract email addresses from text.
- URL Extractor: Extract URLs from text.
- Phone Number Extractor: Extracts North American phone numbers from text and returns them with unified formatting. All the numbers extracted will be valid under the North American Numbering Plan, which means they can be from the US, from Canada, or from certain Caribbean countries.
- Person Extractor: Extract person names from text.
- Person Extractor (Spanish): Extract person names from text in Spanish.
- Price Extractor: Extract prices in different currencies from text. The number and currency are returned separately for more convenient parsing.
- Company Extractor: Extract company and organization entities from text.
- Company Extractor (Spanish) : Extract company and organization names from text in Spanish.
- Location Extractor: Extract location names from text.
- Location Extractor (Spanish): Extract location names from text in Spanish.
- Laptop Features Extractor: Extract features form laptop descriptions.
Check out and explore the public models here. New public models will be continuously added and improved by the MonkeyLearn team.