Fill Ratio

Fill-ratio is maintained by Michael Larson (mjlarson@nbi.ku.dk).

Overview

The fill-ratio is an algorithm that is meant to distinguish track-like events from point-like events. The main idea is that, given a vertex hypothesis (e.g. from c-first), hits should be distributed about that position in some manner. This algorithm first calculates the mean and width of that distribution. Two spheres are drawn around the vertex position with radii of a few times the RMS and a few times the mean that were previously calculated. For each sphere the number of hit OMs are counted contained therein, and the total number of OMs contained therein. The fill-ratios for each sphere, as well as the mean and RMS, and the sphere radii are filled into a frame object, which is then pushed into the frame.

Algorithm

Fill-ratio is a project introduced in the AMANDA days to separate out tracks from cascades at a few hundred GeV. The concept is relatively straightforward and can be described in roughly four steps:

  1. Pick a vertex and hit series. This will define all later steps, with a moderately good vertex giving separation between tracks and cascades. The hit series can be cleaned or uncleaned: In general, a cleaned hit series will benefit the variable, although you will need to optimize this for your own analysis.

  2. Define a “radius” to use. Fill-ratio automatically attempts to use five different radii: a. “FillRatio”: The radius is defined by the distance R between the vertex and each hit:

    \[r_{rms} = \sqrt{\left(\sum{\vec{R}\cdot\vec{r}}\right) - \left(\sum{\vec{R}}\right)^2}\]

    1. “FillRatioFromMean”: The radius is defined by the average distance between each hit in and vertex

      \[r_{mean} = \frac{1}{N}\sum{\left|\vec{R}\right|}\]

    2. “FillRatioFromMeanPlusRMS”: Use both of the above definitions to pick the radius:

      \[r_{mean\;rms} = r_{mean} + r_{rms}\]

    3. “FillRatioFromEnergy”: Define an “SPE radius” based on the reconstructed energy of the given particle. This is an OLD calculation and should be considered unreliable/deprecated!

      \[r_{Energy} = -0.6441 + 36.9 * Log_{10}(E)\]

    4. “FillRatioFromNCh”: Define an “SPE radius” based on the number of hit DOMs in the given hit series. First the nChannel is converted to an “energy”. The energy is then passed to the same calculation as the “FillRatioFromEnergy”. This is an OLD calculation and should be considered unreliable/deprecated!

      \[E_{nCh} = 0.084 + 1.83 Log_{10}(nCh)\]

  3. Once a radius is calculated, find all DOMs within a sphere centered on the given reconstructed vertex with radius given by one of the five methods in (2).

  4. Return the fraction of hit DOMs inside the sphere to unhit DOMs in the sphere.

This supposes that the event is likely to be spherically symmetric if it is cascade-like (ie, close to 1) and relatively empty if track-like (ie, close to 0).

I3FillRatioMoudle returns a specialized object, the I3FillRatioInfo. This object is legacy and can be converted to a standard I3MapStringDouble using I3FillRatio2StringDoubleMap. This is also done by default using the included tray segment.

I3FillRatioLite is an updated version of the I3FillRatioModule code and only performs the calculation using the radius defined by the mean distance (ie, option 2.b above). It offers a cleaner interface and should be preferred for new users.

Usage

I3FillRatioModule (C++ Module)

VertexName

Name of the I3Particle in the frame to use as a vertex

ResultName

The name that the resulting I3FillRatioInfo object will take in the frame

RecoPulseName

The name of the recopulses to be used

SphericalRadiusRMS

The radius (in units of the RMS) of the sphere to be used to calculate the fill ratio.

SphericalRadiusMean

The radius (in units of the mean) of the sphere used to calculate the fill ratio.

SphericalRadiusMeanPlusRMS

The radius (in units of the mean plus rms) of the sphere used to calculate the fill ratio.

SphericalRadiusNCh

The radius (in units of the SPE Radius) of the sphere used to define the fill-ratio

AmplitudeWeightingPower

The means and RMSs can be weighted by the charge of the hit. This parameter sets the power for the exponential weighting.

BadDOMList

Name of the BadDOMList in the DetectorStatus frame

BadOMs

A user-defined vector of OMKeys that should not contribute to the calculation of the fill-ratio.

I3FillRatioLite (C++ Module)

Vertex

Name of the particle to use as a vertex

Pulses

Name of the pulse series in the frame

Output

Name of the I3Double to put in the frame

Scale

Factor by which to scale the mean vertex-DOM distance

AmplitudeWeightingPower

Weight hits by the charge to this power

ExcludeDOMs

List of DOMs to exclude from the calculation

BadDOMListName

Name of the dynamic BadDOMList in the DetectorStatus frame

FillRatio (Tray Segment)

VertexName

Name of the I3Particle in the frame to use as a vertex

ResultName

The name that the resulting I3FillRatioInfo object will take in the frame”

MapName

The name that the I3MapStringDouble object will take in the frame”

RecoPulseName

The name of the recopulses to be used

SphericalRadiusRMS

The radius (in units of the RMS) of the sphere to be used to calculate the fill ratio.

SphericalRadiusMean

The radius (in units of the mean) of the sphere used to calculate the fill ratio.

SphericalRadiusMeanPlusRMS

The radius (in units of the mean plus rms) of the sphere used to calculate the fill ratio.

SphericalRadiusNCh

The radius (in units of the SPE Radius) of the sphere used to define the fill-ratio

AmplitudeWeightingPower

The means and RMSs can be weighted by the charge of the hit. This parameter sets the power for the exponential wieghting.

If

the usual python function, makes the segment run conditionally frame by frame.

See Also