Saffron Technology just may have struck the mother lode of gold in data mining: Securing a patent for technology it has invented that the Cary-based company claims can help predict future events.
On Thursday, Saffron announced it had been awarded a Patent for a means of turning so-called “big data” – massive amounts of information gathered from a wide variety of not necessarily related sources – to forecast what might happen.
The patent is titled:
“Event-Based Anticipation Systems, Methods and Computer Program Products for Associative Memories Wherein Times of Future Events Occurring Are Predicted.”
A drawing on a napkin displayed at Saffron’s website breaks down the Saffron analysis to a simple concept: Data in, analysis, actionable intelligence out. Of course, the process is much more complicated that that, and the U.S. Patent Office in its examination of the Saffron “secret sauce” for analytics granted protection.
“Big data” analytics is huge business these days with companies such as Cary-based SAS, IBM and others investing heavily in the field while clients worldwide are paying billions for the actionable intelligence that analysis produces. Research firm IDC estimates that big data services will be a $16.9 billion market by 2015, up from $3.2 billion in 2010.
Saffron sees a big target for the technology’s use: Wall Street.
“Big-data analytics offers a breakthrough for fund managers and banks, enabling them to anticipate the ups and downs of financial markets,” Saffron says. “Such situational awareness is based on a rich understanding of context including patterns from news sources, social media and other proprietary data. The benefits and collateral damage caused by mergers and acquisitions, business expansion and contraction for example, are widely covered in news sources. Yet more advanced analytics, which includes considering all the information available, is required to understand and exploit such events.”
Saffron, a 13-year-old company with many connections to the defense industry, focuses on analytics software that it says can provide “decision support, predictive [customer relationship management], risk intelligence, security threat intelligence, market intelligence and more.”
The patented technology enables Saffron to use predictive analytics to associate “events to each other in order to anticipate yet other, future events.”
In explaining its “Natural Intelligence Platform,” the basis of its analytics capabilities, Saffron notes:
“[D]ata actually becomes more than … just data. It’s more than information, more than knowledge, more than intelligence. It’s Experience. That is what your data becomes after it’s been ingested into SaffronMemoryBase, and transformed by way of Associative Memory technology. Then, an entire new world opens up to your decision-makers – a world in which they can discover far more useful answers faster than they ever could before with traditional data analytics tools or search / analytics engines.
“In short, its ability to apply your Experience and know more than before about any situation is what’s different about Saffron Technology’s data analytics solutions.”
“Semantics, Statistics and Context”
In a statement about the patent, Saffron Chief Executive Officer Manual Aparicio, a former IBM scientist who co-founded the firm along with another ex-IBMer Jim Fleming, described the advance as enabling the use of data across “connections and counts in context.”
That’s the “Three Cs” of Saffron technology. Saffron also says it integrates “semantics, statistics and context” through its analytics process.
“In entity analytics a mere association between two people can be irrelevant unless you also know the context: how are two people connected. John’s interactions with Fred might be only about golf and never about business. The same is true of events. Knowing that one event tends to follow another event – in general – can be misleading without also knowing when one event is correlated with another, in context of other more complex conditions.
“By introducing this deeper understanding of the event-event context we more accurately anticipate what might happen next and can plan accordingly.”
Cited by research firm Gartner recently as a “new breed of vendor” focused on “hybrid data and analytics,” Saffron says its technology offers numerous strengths to clients:
- “Reduce time to know: 500,000 times faster than reading.
- “Decision support: Up to 10X improvement in operational decision making.
- “Similarity analysis and Nearest neighbour reasoning.
- “Unify Data Sources: Handle internal and external data sources.
- “Unstructured data: Process text, customer records, news, social media etc.
- “Combine Statistics & Semantics: Saffron handles both sorts of analytics.
- “Anticipation: Predict future events using pattern recognition.”
“Associated Memory Analytics”
Saffron integrates what it calls “association memories,” or Associated Memory Analytics, noting that they can be crucial “from remembering the past to anticipating the future.”
In a blog, Saffron talked more about the importance of bringing context to data to better understand the meaning of the data being crunched.
“Rich context is the most significant factor in understanding how to use data in an actionable sense – allowing us to move beyond causal relationships between event A and event B.
“By unlocking a deep understanding of context, Associative Memories help us to better anticipate what could happen, and plan accordingly.”
How Saffron Technology Works
A listing from Saffron describes the analysis process and capabilities:
- Ingests and “reads” data from DataBases (both structured and unstructured data)
- Observes all associations amongst entities and attributes and their frequencies
- Attributes are all words other than entities less words such as “a”, “in”, “the”, etc
- Keeps counts of number of times it observes an association — strength of the link
- Can manage at least a billion entities, can be millions of attributes for each entity
- Nearly infinite scalability — just add more machines
- Operates in real time as new data is streamed to it — not batch processing
- Not traditional search — go directly to entities in question
- No required fixed schema or metadata, but will use whatever is there
- Query response time in sub-seconds to seconds independent of MemoryBase size
- Can provide user access control at the association (link) level
- Runs on standard PC Servers
Saffron clients include research, financial services, risk management, supply chain, national security and business development clients.