|
Mining Knowledge
Now &Then goes beyond data mining to find
and assemble the knowledge specifically relevant to a Query.
At its basic level, Data mining might
retrieve your weight and the number of calories you ate on a specific date, then graph your calories and weight trend over time.
Data mining might also retrieve your home address in 2006. In contrast, Knowledge
mining, "understands" and retrieves journal
information that might help determine why you gained 10 pounds, or explain
why you moved to
your 2006 address, lived
there, or moved away.
Now &Then extracts knowledge, identifies relevant, but forgotten
information, and reveals meaningful relationships in journal information
that may not be superficially evident―at least not apparent based on the key words in
either your Query or the retrieved information!
Now &Then employs knowledge mining methods and techniques that
classify, model, analyze, "clean" and "reduce" data so that useful knowledge
can be discovered. Using rules, and deductive, and Inductive analysis,
Now
& Then
discovers and
extracts non-surface information from the "raw" content of your journal.
Now &Then can do a superior job learning and finding your information
because it does not search through a vast amount
of unrelated information. It focuses completely on you and your information.
It learns how you describe your world of activities, experiences,
reflections, perceptions, and philosophy―using machine learning it grows and changes with you
throughout your lifetime.
Machine Learning
To continually improve Now &Then's understanding of your world, machine learning employs user
rules, user ontologies, system training, and user feedback. Machine learning is
also based on system introspection and Journal use
patterns.
Now &Then knows about your world
because it is trainable, customizable, and
learns automatically to "understand" facts about your life, your vocabulary, and how you use
your journal. As Now &Then learns more about how you
perceive the world, it makes better decisions about how to store, retrieve, and display
your personal information.
Now &Then power and usefulness is continuously enhanced through machine
training and personalization provided by you. It uses
semantic Indexing, multi-dimensional term scaling, knowledge & concept
classification, (Bayesian) query reduction, and relevance
feedback to understand your interests and worldview―even as your experience,
reflections, and philosophy grow and change.
Now &Then understands your worldview because it searches only your journal
information―it does not search through the endless amount of Internet and
online material―information that has nothing to do with your life record. With
every journal entry, Now &Then gets to know you, its user―a
single, individual who uses words in your own, unique way to describe your activities, experiences and perceptions.
see also: Machine
Learning
home
|
