Glossary of statistical terms & studies used in my proprietary trading systems.
A
equal to mean 2". The alternative hypothesis are "mean 1 is not equal to mean 2" and mean 1 is greater than/less
than mean 2."
Analysis of variance - A method for determining the proportion of the variation in a sample that is explained by
one or more classification factors.
B
Bar graph - A graph representing values as bars, where the length of the bars
reflect the values.
Box plot - A graphical representation that depicts the distribution of a response
for different levels of a
categorical predictor.
C
Central probability - The probability that a value of a random variable with a
distribution symmetric about zero
(like the normal distribution) would be closer to zero that a given value (x).
Classical analysis - A method that assumes knowledge of the distribution underlying
the sample. Classical analysis
uses statistical tests based on estimates of the parameters of the underlying population,
such as the mean and the
variance. Such tests are called parametric tests.
Confidence interval - A range of values that have at least the specified
probability of containing the parameter value
being estimated.
Correlation - A measure of the extent to which two variables tend to be associated,
vary, or occur together.
Correlation can be positive or negative and is scaled to lie between -1 and +1. If two
variables are positively
correlated, their values tend to increase and/or decrease together. If two samples are
negatively correlated, the
values of one increase while the values of the other decrease. In addition, correlation
can be linear or non-linear.
If two samples are linearly correlated, there values tend to fall along a straight line
when plotted against one another.
If two samples are nonlinearly correlated, their values tend to follow a nonlinear
pattern.
Covariance - A measure of the extent to which two variables tend to vary together
linearly. The covariance
between two variables is their linear correlation times the product of each of their
standard deviations.
D
Degrees of Freedom - 1) A parameter of certain families of probability
distributions (e.g., the chi squared
distribution); 2) a measure of the number of dimensions in statistical data or model
structure. The degrees of
freedom of the structure are the number of quantities that can vary independently in that
structure.
Descriptive statistics - Statistical measures of data that reflect properties of
the data and provide a means of
assessing the general nature of the underlying distribution. Examples include mean,
standard deviation, range.
Dispersion - The extent to which data are spread out rather than clustered close to
their mean or median.
Distribution - The relative frequency in which values of a variable in a data
sample or in a population are
distributed across the range of possible values.
H
Histogram - Represents the frequency distribution of values of a variable by
rectangular bars. The width of the
bars represent classifications or ranges of values of the variable. The heights of the
bars represent the frequencies
of values falling into the corresponding classification or range.
Hypothesis - An assertion or conjecture about the distributions of one or more
variables
K
or negative. The standard normal distribution has a kurtosis of zero and is the standard against which the kurtosis
values for other distributions are measured. The larger the kurtosis, the more peaked the distribution and the more
frequent the outliers. The smaller the kurtosis the flatter the distribution.
M
Mean - The sum of values in a sample, divided by the count (total number of
values).
The mean us similar to the median in that it estimates the "center" of the
distribution of data.
Median - An estimate of the center of the distribution of the data. It is the value
above and below which lie an equal
number of data values when the data are arranged in increasing or decreasing order. If the
count of the sample is odd,
the median is the middle value. If the count is even, the median is the average of the two
middle values.
Model - An equation relating responses and predictors. A statistical model contains
two parts: a model for the signal
and a model for the noise, or error, associated with the signal. The model for the signal
is an equation describing how
the mean of the responses depends on certain predictor variables. The model for the noise,
or error, describes the
distribution of the deviations of responses from the signal.
N
Null hypothesis - A hypothesis that asserts either the equivalence of two unknown
quantities or the positive statement
of a condition (for example, "mean 1 is equal to mean 2" or "the population
is normal"). It is used in hypothesis testing
as the default hypothesis in contrast to the alternative hypothesis.
O
Outlier - An extreme value in a data sample that is far from the central cluster of
values in the sample. An outlier is
extraordinary either because it is "far away" from the other values or because
it fails to conform to the patter of the
other values.
P
Parameters - Quantities that characterize or are functions of the distribution of a
population, such as mean and
variance.
Parametric tests - Statistical tests based on knowledge of the parameter form of
the distribution of the underlying
population. Parameter tests are used in classical analysis.
Population mean - The unknown mean of the underlying population from which data
values are drawn.
Predicted values - Values predicted by a fitted model.
Probability distribution - A representation of the probabilities associated with
values of a random variable.
R
Random variable - A variable whose value is determined by the outcome of an event
or experiment.
Range - The difference between the maximum and the minimum values in a group of
data.
Regression Analysis - Analysis that is used to depict the relation between
dependent and independent variables
derived from a study. The goal of regression is to find a mathematical model that best
explains how changes in the
independent predictor variables affect the dependent response variables. A fitted
regression model estimates the
relation between predictors and responses and may be used to forecast or predict values of
the responses.
S
Skewed distribution - A distribution that is not symmetrical about its mean, but
instead has values concentrated
on one side or the other of the mean. If the distribution is concentrated to the left of
the mean and is more spread out
to the right of the mean (when graphed), the distribution is positively skewed. If the
reverse is true, the distribution is
negatively skewed
Skewness - A measure of the degree of asymmetry of the data around the sample
mean. If the data is distributed
symmetrically around the mean, the skewness is 0.0. From a graphical point of view,
negative skewness indicates that
the data are more spread out to the left of the mean than to the right of it. The reverse
is true for positive skewness.
Standard deviation - A measure of the average deviation of data values from the
sample mean. The standard
deviation is equal to the square root of the variance.
U
Unimodal - A distribution whose histogram has just one "hump", as in a
bell-shaped curve. (Bimodal distribution
would have two "humps" )
V
Variable - A measurement or attribute of each member of a sample or population, for
example, weight, height, age.
Variance - The average of the squared differences between the values of a variable
and the mean. The variance
estimates the spread of a distribution.
W
Weighted - In order to adjust for the fact that some data are less valuable or less
reliable than other data, a statistical
analysis may down weight the less reliable points for certain computations.
X
X-Y graph - A visual representation that depicts the relationship between a
continuous predictor and a continuous
response; can be a scatter plot or a line plot.
Membership
to Floyd Upperman & Associates is like having
your own personal CTA! Sign
up today!
I'll do everything I can to help
you learn my
unique and powerful trading methods, which
combine technical and fundamental analysis!
My personal guarantee is a 200%
effort...ALWAYS!
NO TRADING SYSTEM CAN
GUARANTEE PROFITS
No guarantees can be made for
success. Past performance is not a guarantee of future profits.
Futures trading is NOT my only means of income. I also
invest in stocks, real estate as well as
generate income from
other businesses. I have both winning trades and losing trades.
I trade professionally, but not daily. I wait for what I believe to be
ideal trading opportunities.
Very Respectfully,
Floyd W. Upperman Jr.
Click here to get back to the front of this
web-site...
Email comments or questions to floyd@upperman.com
F L O Y D U P P E R M A N & A
S S O C I A T E S
Copyright © 1997 - 2001 by Floyd Upperman & Associates. All rights reserved.
Revised:
30 Jul 2008 11:39:30 -0700.
Futures Trading is a risky business.
Do not trade with money you
can not afford to lose.
Our Disclaimers & Legal Notes.
Only those who will
risk going too far can possibly find out how far one can go.
T. S. Eliot
American-British
Poet