Comparing a new algorithm with the classic methods for estimating the number of factors

Henry, Ronald C.; Park, Eun Sug; Spiegelman, Clifford H.

doi:10.1016/s0169-7439(99)00015-5

Cited by 85 publications

(44 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The number of latent variables for each PLS model was determined by using the knee-finding algorithm in PLS Toolbox where a drop known as the "knee" is located in the scree plot (Henry et al, 1999).…”

Section: Pls Model For Prediction Of Maturation Indices Three Partialmentioning

confidence: 99%

Measuring Cheese Maturation with the Fluorescence Fingerprint

Kokawa

Ikegami

Chiba

et al. 2015

FSTR

View full text Add to dashboard Cite

show abstract

Section: Pls Model For Prediction Of Maturation Indices Three Partialmentioning

confidence: 99%

Measuring Cheese Maturation with the Fluorescence Fingerprint

Kokawa

Ikegami

Chiba

et al. 2015

FSTR

View full text Add to dashboard Cite

show abstract

“…UNMIX solves eqs 2 and 3 by using a principle component analysis (PCA) approach to reduce the number of dimensions in the space to the number of factors that produce the data, followed by an unique "edge detection" technique to identify "edges" defined by the data points in the space of reduced dimension (e.g., Figures 1 and 3). The number of factors is estimated by the NUMFACT algorithm in advance 127 , which reports the R 2 and signal-to-noise (S/N) ratio associated with the first N principle components (PCs) in the data matrix. The number of factors should coincide with the number of PCs with S/N ratio Ͼ 2.…”

Section: Weaknessesmentioning

confidence: 99%

Source Apportionment: Findings from the U.S. Supersites Program

Watson

Chen

Chow

et al. 2008

Journal of the Air & Waste Management Association

222

132

View full text Add to dashboard Cite

Receptor models are used to identify and quantify source contributions to particulate matter and volatile organic compounds based on measurements of many chemical components at receptor sites. These components are selected based on their consistent appearance in some source types and their absence in others. UNMIX, positive matrix factorization (PMF), and effective variance are different solutions to the chemical mass balance (CMB) receptor model equations and are implemented on available software. In their more general form, the CMB equations allow spatial, temporal, transport, and particle size profiles to be combined with chemical source profiles for improved source resolution. Although UNMIX and PMF do not use source profiles explicitly as input data, they still require measured profiles to justify their derived source factors. The U.S. Supersites Program provided advanced datasets to apply these CMB solutions in different urban areas. Still lacking are better characterization of source emissions, new methods to estimate profile changes between source and receptor, and systematic sensitivity tests of deviations from receptor model assumptions. INTRODUCTIONReceptor-oriented source apportionment models infer source contributions and atmospheric processes from air quality measurements. Receptor models complement, rather than replace, source-oriented dispersion and chemical transformation models that begin with source emission rates to estimate ambient concentrations. 1,2 Source and receptor models are mathematical representations of reality, requiring simplifying assumptions that create uncertainty. Applying both types of models to the same situation allows them to be improved when their results diverge and lends confidence to their results when they agree. 3

show abstract

“…While some receptor modeling studies have been previously conducted in the Cincinnati area (Mukerjee and Biswas, 1992;Mukerjee and Biswas, 1993;Shenoi, 1990), detailed local source profiles are not available for PM 2.5 constituents. The UNMIX, multivariate model (Henry, 1994(Henry, , 2003Henry et al, 1999) was adopted in this study to derive factors which presumably can be attributed to emission source categories. To our knowledge this is the first application of a receptor modeling study of PM 2.5 constituents in the Greater Cincinnati area.…”

Section: Unmix Model Calculationsmentioning

confidence: 99%

“…The first step adopted is applying 'NUMFACT' (Henry et al, 1999) to determine the number of influencing sources. This is analogous to factor analysis methods that establish the number of factors (or sources), but with different criteria being invoked.…”

Section: Unmix Model Calculationsmentioning

confidence: 99%

UNMIX modeling of ambient PM2.5 near an interstate highway in Cincinnati, OH, USA

McDonald

Martuzevičius

et al. 2006

Atmospheric Environment

View full text Add to dashboard Cite

The "Cincinnati Childhood Allergy and Air Pollution Study (CCAAPS)" is underway to determine if infants who are exposed to diesel engine exhaust particles are at an increased risk for atopy and atopic respiratory disorders, and to determine if this effect is magnified in a genetically at risk population. In support of this study, a methodology has been developed to allocate local traffic source contributions to ambient PM 2.5 in the Cincinnati airshed. As a first step towards this allocation, UNMIX was used to generate factors for ambient PM 2.5 at two sites near at interstate highway. Procedures adopted to collect, analyze and prepare the data sets to run UNMIX are described. The factors attributed to traffic sources were similar for the two sites. These factors were also similar to locally measured truck engine-exhaust enriched ambient profiles. The temporal variation of the factors was analyzed with clear differences observed between factors attributed to traffic sources and combustion-related regional secondary sources.

show abstract

Comparing a new algorithm with the classic methods for estimating the number of factors

Cited by 85 publications

References 10 publications

Measuring Cheese Maturation with the Fluorescence Fingerprint

Measuring Cheese Maturation with the Fluorescence Fingerprint

Source Apportionment: Findings from the U.S. Supersites Program

UNMIX modeling of ambient PM2.5 near an interstate highway in Cincinnati, OH, USA

Contact Info

Product

Resources

About