Defining Decision Contours
To categorize a new sample external to the training set, apply the previous distance metric to find the transition probabilities between and all samples in, with a normalisation. Secondly apply the last two lines of the 3-line algorithm to get cluster, and conditional category probabilities.
Finally we have
Parameter must be kept under close supervision since, as it is increased from zero, increasing numbers of features, in the category probability space, snap into focus at certain critical thresholds.
Read more about this topic: Information Bottleneck Method
Famous quotes containing the words defining, decision and/or contours:
“The U.S. is becoming an increasingly fatherless society. A generation ago, an American child could reasonably expect to grow up with his or her father. Today an American child can reasonably expect not to. Fatherlessness is now approaching a rough parity with fatherhood as a defining feature of American childhood.”
—David Blankenhorn (20th century)
“The women of my mother’s generation had, in the main, only one decision to make about their lives: who they would marry. From that, so much else followed: where they would live, in what sort of conditions, whether they would be happy or sad or, so often, a bit of both. There were roles and there were rules.”
—Anna Quindlen (20th century)
“Pregnant women! They had that weird frisson, an aura of magic that combined awkwardly with an earthy sense of duty. Mundane, because they were nothing unique on the suburban streets; ethereal because their attention was ever somewhere else. Whatever you said was trivial. And they had that preciousness which they imposed wherever they went, compelling attention, constantly reminding you that they carried the future inside, its contours already drawn, but veiled, private, an inner secret.”
—Ruth Morgan (1920–1978)