Planet Data Analytics
In the 21st century, we are swimming in a virtual sea of unstructured information: emails, documents, forms, and the like make up 85% of the information companies have to manage.
Planet Data Analytics is a unique Latent Semantic Indexing (LSI) based technology that was designed to address the challenges of unstructured information, to make computers search and process this information in a more human-like manner.
Formerly only available to elite government offices and the Intelligence community, the conceptual processing power of Planet Data Analytics is now available to commercial markets.
Latent Semantic Indexing (LSI)
Latent Semantic Indexing (LSI) was originally developed by Bell Laboratories as a way to express conceptual text relationships as mathematics.
Using LSI and a number of derivative software products, Planet Data Analytics technology is a completely different way to address searching, sorting, classifying, and retrieving. There are a number of benefits:
- Information classification and retrieval are based on concepts, not just keywords Conceptual Search is extremely error-resistant – a few misspellings are irrelevant when understanding concepts
- Conceptual Search is multi-lingual: the same mathematic processes applied to English work across languages –the system simultaneously understands documents without translation
- Unique derivative products are possible: automatic categorization, summarization, instant context, social network analysis are just a few byproducts of this technology
Logical - because it is Concept-Based
Humans think in concepts, which is how our technology works. By working with key concepts instead of key words, we transcend word challenges like synonyms and polysemy (words whose meanings change by how they’re used). We also transcend the language barrier – in fact, Planet Data Analytics is fluent in 17 languages today. Finally, working with concepts makes our technology very resistant to errors and misspellings.
Easy - because it uses Machine Learning
Our technology incorporates powerful machine learning and self-training. As it builds an index of information, our technology is continually learning. It can learn by example or entirely on its own. Planet Data Analytics is a better way to sort, process, and retrieve unstructured information.
Planet Data Analytics is a better way to sort, process, and retrieve unstructured information.
A better way to Sort
A simple collection of “example” documents is all Planet Data Analytics needs to train itself on how to best sort unstructured information like emails, documents, incoming correspondence. It will recognize documents by what they are saying, rather than visual cues, even if it has never “seen” that kind of document before.
A better way to Process
Because Planet Data Analytics understands concepts, not merely words, it can automatically create a summary of a long document. It can create “pop-up” context to illustrate the meaning of unfamiliar words based on what has learned from a set of documents.
A better way to Retrieve
Planet Data Analytics will find related information – concepts – that you haven’t yet thought about. It isn’t stumped by synonyms or word errors. It even searches across languages without requiring translation.
How Planet Data Analytics Delivers Value to its Customers
In the challenging landscape of search, classification, and retrieval products, Planet Data Analytics’s customers see value from our products in two ways: either through dramatic cost savings, or from breakthrough products based on our technology.
Cost Savings:
- A litigation support company sees workload reductions of 65% or more using conceptual search vs. multiple keyword passes, while catching expensive coding errors.
- A government agency sees a way to cut $120,000 out of their annual budget to handle incoming email and outside correspondence.
- An opinion research firm sees a way to reduce the cost of delivering survey results by 45%.
New Products:
- An entirely new information product based on genome-mapping for a genomics research firm.
What is Conceptual Search Technology? - Unlike traditional keyword search engines, conceptual searching is based on how and where ideas and concepts intercept with similar ideas and concepts in a document collection. This is mathematically based - there are no word lists or dictionaries, or linguistic techniques (like sentence structures). The technology is self-training, “learning” all it requires from material being processed. Not only can it identify, classify, and sort information rapidly, it will also find relevant documents that other techniques routinely miss.
Conceptual Search - Feature: Run searches based on how and where ideas and concepts intercept with similar ideas and concepts in a document collection. Enter blocks of relevant text or the contents of entire documents as search criteria. Benefits: Find more of the relevant documents and less junk. Do it faster and more accurately than by constructing multiple elaborate keyword searches. Avoid the pitfalls of OCR errors affecting search results.
Integrated Keyword and Conceptual Search - Feature: Combine keywords and concepts in one integrated search. Benefits: Eliminate the issue of a single keyword having multiple meanings. Find the documents you intend the search to return and avoid false positive hits.
