SweetSpotSimilaritypublic class SweetSpotSimilarity extends DefaultSimilarity | A similarity with a lengthNorm that provides for a "platuea" of
equally good lengths, and tf helper functions.
For lengthNorm, A global min/max can be specified to define the
platuea of lengths that should all have a norm of 1.0.
Below the min, and above the max the lengthNorm drops off in a
sqrt function.
A per field min/max can be specified if different fields have
different sweet spots.
For tf, baselineTf and hyperbolicTf functions are provided, which
subclasses can choose between.
|
| Fields Summary |
|---|
| private int | ln_min | | private int | ln_max | | private float | ln_steep | | private Map | ln_mins | | private Map | ln_maxs | | private Map | ln_steeps | | private float | tf_base | | private float | tf_min | | private float | tf_hyper_min | | private float | tf_hyper_max | | private double | tf_hyper_base | | private float | tf_hyper_xoffset |
| Constructors Summary |
|---|
public SweetSpotSimilarity()
super();
|
| Methods Summary |
|---|
| public float | baselineTf(float freq)Implimented as:
(x <= min) ? base : sqrt(x+(base**2)-min)
...but with a special case check for 0.
This degrates to sqrt(x) when min and base are both 0
if (0.0f == freq) return 0.0f;
return (freq <= tf_min)
? tf_base
: (float)Math.sqrt(freq + (tf_base * tf_base) - tf_min);
| | public float | hyperbolicTf(float freq)Uses a hyperbolic tangent function that allows for a hard max...
tf(x)=min+(max-min)/2*(((base**(x-xoffset)-base**-(x-xoffset))/(base**(x-xoffset)+base**-(x-xoffset)))+1)
This code is provided as a convincience for subclasses that want
to use a hyperbolic tf function.
if (0.0f == freq) return 0.0f;
final float min = tf_hyper_min;
final float max = tf_hyper_max;
final double base = tf_hyper_base;
final float xoffset = tf_hyper_xoffset;
final double x = (double)(freq - xoffset);
final float result = min +
(float)(
(max-min) / 2.0f
*
(
( ( Math.pow(base,x) - Math.pow(base,-x) )
/ ( Math.pow(base,x) + Math.pow(base,-x) )
)
+ 1.0d
)
);
return Float.isNaN(result) ? max : result;
| | public float | lengthNorm(java.lang.String fieldName, int numTerms)Implimented as:
1/sqrt( steepness * (abs(x-min) + abs(x-max) - (max-min)) + 1 )
This degrades to 1/sqrt(x) when min and max are both 1 and
steepness is 0.5
:TODO: potential optimiation is to just flat out return 1.0f if numTerms
is between min and max.
int l = ln_min;
int h = ln_max;
float s = ln_steep;
if (ln_mins.containsKey(fieldName)) {
l = ((Number)ln_mins.get(fieldName)).intValue();
}
if (ln_maxs.containsKey(fieldName)) {
h = ((Number)ln_maxs.get(fieldName)).intValue();
}
if (ln_steeps.containsKey(fieldName)) {
s = ((Number)ln_steeps.get(fieldName)).floatValue();
}
return (float)
(1.0f /
Math.sqrt
(
(
s *
(float)(Math.abs(numTerms - l) + Math.abs(numTerms - h) - (h-l))
)
+ 1.0f
)
);
| | public void | setBaselineTfFactors(float base, float min)Sets the baseline and minimum function variables for baselineTf
tf_min = min;
tf_base = base;
| | public void | setHyperbolicTfFactors(float min, float max, double base, float xoffset)Sets the function variables for the hyperbolicTf functions
tf_hyper_min = min;
tf_hyper_max = max;
tf_hyper_base = base;
tf_hyper_xoffset = xoffset;
| | public void | setLengthNormFactors(int min, int max, float steepness)Sets the default function variables used by lengthNorm when no field
specifc variables have been set.
this.ln_min = min;
this.ln_max = max;
this.ln_steep = steepness;
| | public void | setLengthNormFactors(java.lang.String field, int min, int max, float steepness)Sets the function variables used by lengthNorm for a specific named field
ln_mins.put(field, new Integer(min));
ln_maxs.put(field, new Integer(max));
ln_steeps.put(field, new Float(steepness));
| | public float | tf(int freq)Delegates to baselineTf
return baselineTf(freq);
|
|
