Abstract
We have many kinds of data of time series such as stock prices.
In the previous research we expressed global trends via
a representative value on the fuzzy intervals in the temporal axis,
local features via the position of large difference
between the original data and
the data generated from the global trends, and
oscillation via the standard deviation and the number of oscillations.
However, the method to count the number of oscillations
is not natural for some data.
We propose a method to count the number of oscillations
by the angles with the lines that link the adjoining data.
The greater the sum of the angles is, the less
the number of oscillations is.
The greater the difference of the angles is,
the less the number of oscillations is.
And we also consider the scale of graph,
which effects the angles of the lines.
