The Formula tab is where you select the numeric and logical functions that are allowed to participate in your model development. In general, it is not wise to simply select all possible operations, because that will hamper the ability of the optimizer to find the correct ones for your data. If possible, you should select only the operations you feel are either possible or appropriate. Click here for tips on how to select the appropriate operations.
Here is a list of the possible operations by category, along with a short description.
Operation

Description

Special Comments

Arithmetic



x+y

x plus y


xy

x minus y


x*y

x times y


x/y

x divided by y

if y=0 result is 0

Algebra



sqrt(abs(x))

square root of absolute value of x


x

negative of x


1/x

inverse of x = 1 divided by x

if x=0, result is 0

x^2

x squared = x*x


x^3

x cubed = x*x*x


Abs(x)

absolute value of x


Trigonometry and Transcendental



x^y

x raised to the power y

result is 0 if x<=0, y>=20, or y<=20

exp(x)

base of natural logs e raised to the power x

result is exp(50) if x>50

10^x

10 raised to the power x

result is 0 if x>=20

log(x)

logarithm of x

result is 0 if x<=0

ln(x)

natural logarithm of x

result is 0 if x<=0

sin(x)

sine of x

x is in radians

cos(x)

cosine of x

x is in radians

Neural



neuron2(a,b,c,d)

tanh(a*b + b*c)

tanh is hyperbolic tangent

neuron3(a,b,c,d,e,f)

tanh(a*b + c*d + e*f)

tanh is hyperbolic tangent

neuron4(a,b,c,d,e,f,g,h)

tanh(a*b + c*d + e*f + g*h)

tanh is hyperbolic tangent

tanh(x)

hyperbolic tangent of x

activation function used in neural nets

sigmoid(x)

1/(1+exp(x))

activation function used in neural nets

Boolean



x and y

true(1) if both x and y are true, else false(0)

0=false, nonzero=true

x or y

true(1) if either x or y is true, else false(0)

0=false, nonzero=true

not x

true(1) if x is false, false(0) if x is true

0=false, nonzero=true

if x then y else z

result is y if x is true, otherwise result is z

0=false, nonzero=true

Relational



x<y

true(1) if x is less than y, else false(0)


x>y

true(1) if x is greater than y, else false(0)


x>=y

true(1) if x greater than or = y, else false(0)


x<=y

true(1) if x less than or = y, else false(0)


Polynomials



a^2 + b^2

a squared plus b squared


a^2 – b^2

a squared minus b squared


a*b + b*c + d*e

the sum of three products


Statistical



min(x,y)

the minimum of x and y


max(x,y)

the maximum of x and y


avg(x,y)

the average of x and y


Technical Indicators



Momentum

current value  a previous value


Rate of change

current value / a previous value


Rate of change %

rate of change * 100


% change

100*( current value  previous value) / previous value


Velocity

(current value – value N periods ago) / N


Acceleration

(current velocity – velocity N periods ago) / N


Min value

minimum value over N periods


Max value

maximum value over N periods


Spread

value1 – value2


Spread %

100*(value1 – value2)/value1


Relative strength

value1 / value2


Efficiency

click here for definition


Simple moving avg

sum of last N values / N


Exponential moving avg

click here for definition


Slope

click here for definition


Lag

a previous value


Click here for more details on how to create technical indicators such as momentum, velocity, slope, etc, in ChaosHunter.
