Log-normal Distributions in Gamma-Ray Burst Time Histories

Li, Hui; Fenimore, E. E.

doi:10.1086/310275

Cited by 65 publications

(92 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Li-Fenimore algorithm (LFA) operated by binning the data and then labeling as a candidate peak each bin that had more counts than its immediate neighbors Li & Fenimore (1996). A search was then conducted for each candidate peak to determine if the counts for non-immediate neighbors (more than one bin away) continued to diminish according to a given formula.…”

Section: Related Work On Peak Identificationmentioning

confidence: 99%

“…RAPID also differs from other binning techniques for burst detection or peak identification (Li & Fenimore 1996;Guidorzi 2015;Karako-Argaman et al 2015) in several key ways. First, all other binning techniques look at only a single value for each bin.…”

Section: Dpg Identification With Rapidmentioning

confidence: 99%

“…However, this cannot tell us which direction the points inside one bin are trending. Additionally, RAPID only needs to make one pass through the data, while LFA (Li & Fenimore 1996) and MEPSA (Guidorzi 2015) perform an initial pass to identify candidates and then another pass to search for patterns.…”

Section: Dpg Identification With Rapidmentioning

confidence: 99%

See 2 more Smart Citations

Detection of dispersed radio pulses: a machine learning approach to candidate identification and classification

Devine¹,

Goševa-Popstojanova²,

McLaughlin³

2016

Mon. Not. R. Astron. Soc.

View full text Add to dashboard Cite

Searching for extraterrestrial, transient signals in astronomical data sets is an active area of current research. However, machine learning techniques are lacking in the literature concerning single-pulse detection. This paper presents a new, two-stage approach for identifying and classifying dispersed pulse groups (DPGs) in single-pulse search output. The first stage identified DPGs and extracted features to characterize them using a new peak identification algorithm which tracks sloping tendencies around local maxima in plots of signal-to-noise ratio vs. dispersion measure. The second stage used supervised machine learning to classify DPGs. We created four benchmark data sets: one unbalanced and three balanced versions using three different imbalance treatments. We empirically evaluated 48 classifiers by training and testing binary and multiclass versions of six machine learning algorithms on each of the four benchmark versions. While each classifier had advantages and disadvantages, all classifiers with imbalance treatments had higher recall values than those with unbalanced data, regardless of the machine learning algorithm used. Based on the benchmarking results, we selected a subset of classifiers to classify the full, unlabelled data set of over 1.5 million DPGs identified in 42,405 observations made by the Green Bank Telescope. Overall, the classifiers using a multiclass ensemble tree learner in combination with two oversampling imbalance treatments were the most efficient; they identified additional known pulsars not in the benchmark data set and provided six potential discoveries, with significantly less false positives than the other classifiers.

show abstract

Section: Related Work On Peak Identificationmentioning

confidence: 99%

Section: Dpg Identification With Rapidmentioning

confidence: 99%

See 1 more Smart Citation

Detection of dispersed radio pulses: a machine learning approach to candidate identification and classification

Devine¹,

Goševa-Popstojanova²,

McLaughlin³

2016

Mon. Not. R. Astron. Soc.

View full text Add to dashboard Cite

show abstract

“…That we used MEPSA instead of other algorithms, such as the popular one proposed by (Li & Fenimore 1996, hereafter LF), was motivated by its much lower false-positive probability (FP; 1-2×10 −5 to be compared with 3-5×10 −3 FP bin −1 ), particularly when the signal drops to background, and by its capability to detect slowly varying, dim peaks. Essentially, MEPSA simultaneously compares various moving intervals (with different lengths) with adjacent bins against a number of thresholds (in units of statistical noise) that must be fulfilled simultaneously to trigger at least one out of 39 different criteria.…”

Section: Real Data Selectionmentioning

confidence: 99%

“…In the case of GRBs a number of papers have investigated this property (McBreen et al 1994;Norris et al 1996). The different peak detection algorithms (Li & Fenimore 1996;Nakar & Piran 2002;Drago & Pagliara 2007;Bhat et al 2012) or techniques (Quilligan et al 2002) yielded as the main result that the GRB ITD is generally well described by a lognormal with mean values < ∼ 1 s, with evidence for a power-law excess at relatively long (Δt > 5-10 s) ITs. These long, rare ITs during which the GRB signal drops to background are often referred to as quiescent times (QTs) and are interpreted as caused by something different from what rules the shorter and more frequent ITs (e.g., Drago et al 2008;Tchekhovskoy & Giannios 2014).…”

Section: Introductionmentioning

confidence: 99%

Gamma-ray burst engines may have no memory

Baldeschi

Guidorzi

2014

A&A

View full text Add to dashboard Cite

Context. A sizeable fraction of gamma-ray burst (GRB) time profiles consist of a temporal sequence of pulses. The nature of this stochastic process carries information on how GRB inner engines work. The so-called interpulse time defines the interval between adjacent pulses, excluding the long quiescence periods during which the signal drops to the background level. It was found by many authors in the past that interpulse times are lognormally distributed, at variance with the exponential case that is expected for a memoryless process. Aims. We investigated whether the simple hypothesis of a temporally uncorrelated sequence of pulses is really to be rejected, as a lognormal distribution necessarily implies. Methods. We selected and analysed a number of multi-peaked CGRO/BATSE GRBs and simulated similar time profiles, with the crucial difference that we assumed exponentially distributed interpulse times, as is expected for a memoryless stationary Poisson process. We then identified peaks in both data sets using a novel peak search algorithm, which is more efficient than others used in the past. Results. We independently confirmed that the observed interpulse time distribution is approximately lognormal. However, we found the same results on the simulated profiles, in spite of the intrinsic exponential distribution. Although intrinsic lognormality cannot be ruled out, this shows that intrinsic interpulse time distribution in real data could still be exponential, while the observed lognormal could be ascribed to the low efficiency of peak search algorithms at short values combined with the limitations of a bin-integrated profile. Conclusions. Our result suggests that GRB engines may emit pulses after the fashion of nuclear radioactive decay, that is, as a memoryless process.

show abstract

Gamma-ray bursts—fireballs and blastwaves

2009

Gamma-Ray Bursts

View full text Add to dashboard Cite

Log-normal Distributions in Gamma-Ray Burst Time Histories

Cited by 65 publications

References 8 publications

Detection of dispersed radio pulses: a machine learning approach to candidate identification and classification

Detection of dispersed radio pulses: a machine learning approach to candidate identification and classification

Gamma-ray burst engines may have no memory

Gamma-ray bursts—fireballs and blastwaves

Contact Info

Product

Resources

About