The exponential increase of product reviews by opinionated users, as well as instant social media comments and blog posts calls for tools that can provide a solution for quickly ascertaining the public’s sentiment and opinion on a wide range of subjects, in any language.
Pangea Sentiment Analysis Tool is a customizable, powerful and efficient tool to automatically extract positive or negative sentiments from written text. It can be fine tuned to detect particular emotions (strong dislike, deep like, fear, anger, disgust, etc.) from unstructured text information. It is API ready for immediate analysis at a sentence level (snippet) or for processing texts and documents by batches. Pangea Sentiment Analysis Tool can also be used for other purposes, such as document sentiment classification and review rating prediction tasks.
How does Pangea Sentiment Analysis tool work?
There are two main approaches to sentiment analysis: the traditional “lexicon-based” and the new “learning approach”.
The traditional approach to sentiment analysis is lexicon-based. Through this approach, the semantic orientation of words in a text is calculated by obtaining word polarities from a pre-defined lexicon. The supervised learning approach uses heavy neural machine learning techniques to create a specific model from a large corpus of documents. This data varies from client to client and may include relevance selection (a prior query to verify if a document belongs to the group and should be analyzed).
A set of sample positive and negative opinions and variations of those opinions makes the training data from which the model is built. Our machine learning techniques achieve more than 80% in sentiment accuracy right off. Customization and human-built linguistic filters in each case add to the expected quality to obtain over 90% accuracy. Sentiment analysis’ difficulty is due to the structure of opinions, often dependent on the domain and sometimes context-sensitive.
Pangea Sentiment Analysis Tool unique hybrid approach to sentiment analysis is based on using not only linguistic and statistical information, but also a set of language-dependent complex semantic rules. These, together with the use of deep neural networks to classify irony and seemingly negative or positive results provide a unique and fast sentiment classification experience.
Approach to sentiment analysis successfully process:
Our linguistic experts have created custom built ontologies for the most significant and widely used data domains (hotels, restaurants, gadgets, etc.). These used to be the basis of sentiment analysis, but now feed deep neural networks for which data has been pre-tagged. Sentiments can be grouped in different ways, according to the client’s preferences.
Here, red means negative, blue neutral and green positive sentiment. These are real samples coming from the combination of our hybrid techniques:
The time of the play was short.
Love the short battery recharge time
Neural detection (context)
Bailey short-circuited me. That’s an artist.
Negatives /Intensifiers (neural detection)
Not bad a cheeseburger for Beijing.
Too quick their service for the price tag.
Small. Light. Portable. 16Mb. Thanks Dad!
S***ng, put its shit together. Now, yes. Worth every dollar.
Privacy & Cookies Policy
Necessary cookies are absolutely essential for the website to function properly. This category only includes cookies that ensures basic functionalities and security features of the website. These cookies do not store any personal information.
Any cookies that may not be particularly necessary for the website to function and is used specifically to collect user personal data via analytics, ads, other embedded contents are termed as non-necessary cookies. It is mandatory to procure user consent prior to running these cookies on your website.