On Grouping for Maximum Homogeneity

Fisher, Walter D.

doi:10.1080/01621459.1958.10501479

Cited by 621 publications

(282 citation statements)

References 3 publications

Supporting

Mentioning

276

Contrasting

Unclassified

Order By: Relevance

“…Since the weights are automatically defined once the subarray configuration has been determined, the original compromise problem reduces to that of finding the aggregation of array elements that optimizes the following cost functions: ∈ is the m-th sub-array index. In order to determine C a reduced set of solutions, called contiguous partitions [10], is considered. Successively, the exploration of the solution space is performed just modifying some elements of the solution called "border elements".…”

Section: Problem Statementmentioning

confidence: 99%

Optimization of difference pattern features in sub-arrayed monopulse antennas through and excitations matching procedure

Rocca

Manica

Benedetti

et al. 2008

2008 IEEE Antennas and Propagation Society International Symposium

View full text Add to dashboard Cite

Section: Problem Statementmentioning

confidence: 99%

Optimization of difference pattern features in sub-arrayed monopulse antennas through and excitations matching procedure

Rocca

Manica

Benedetti

et al. 2008

2008 IEEE Antennas and Propagation Society International Symposium

View full text Add to dashboard Cite

“…Un partitionnement univarié [18] a été utilisé pour regrouper les fruits en classes homogènes de pression de façon optimale en minimisant l'inertie intra-classe (figure 4). Les fruits retenus ont été ceux dont la valeur de la pression était la plus proche du barycentre de la classe à laquelle ils appartiennent.…”

Section: Constitution De Sous-lots Homogènes Reflétant L'hétérogénéitunclassified

Méthodes d’échantillonnage de fruits appliquées à l’évaluation de la qualité de l’abricot

Grotte

Gouble

Reling

et al. 2006

Fruits

View full text Add to dashboard Cite

Sampling methods of fruits applied to the quality characterization of apricot fruits.Abstract --Introduction. Heterogeneity of apricot batches results from the difficulty of objectively appraising the optimal harvest date and, thus, the variability of fruit development traits at harvest time. This heterogeneity is a serious problem when comparing between harvests, varieties, and fruit farming cultivation practice. Three sampling procedures were tested for use prior to the characterization of the physiological traits of apricot batches using destructive or non-destructive physical methods. The objective was not only to reduce the number and cost of the analyses but also to achieve homogeneous samples representative of the harvest batch and of its heterogeneity, and to select comparable samples. Materials and methods. Sampling was grounded on whole fruit compression test values (stress) up to a compression rate that does not damage the fruit. Sample batches were then selected after the statistical treatment (descriptive analysis, univariate clustering) of strain values. The sampling methods were applied (a) to compare physical growth traits of apricot hybrids from 2 years with contrasted weather (normal, heatwave), (b) to study the effect of cultivation practices on some physical properties of seven apricot hybrids and (c) to follow postharvest changes. Firmness, color and specific gravity of the fruit and the flesh were measured. Results and discussion. The selection of median samples permitted different harvest batches to be distinguished and the identification of the weather effect on fruit traits (delayed changes to fruit color and flesh specific gravity, less intensely colored flesh). Univariate sampling based on fruit firmness proved to be adequate for characterizing the heterogeneity of harvest batches and for comparing the physical properties of several apricot hybrids. The selection of batches with comparable variability of firmness permitted the evaluation of the effect of repeated non-destructive physical measurements on apricot postharvest changes in simple batches of fruit.France / Prunus armeniaca / fruits / sampling / methods / physicochemical properties / firmness / color / density Méthodes d'échantillonnage de fruits appliquées à l'évaluation de la qualité de l'abricot.Résumé --Introduction. L'hétérogénéité de maturité des récoltes d'abricots résulte de la difficulté d'évaluer objectivement le stade optimal de récolte et donc la variabilité du stade de déve-loppement des fruits au moment de la cueillette. Cette hétérogénéité pose un problème lorsqu'il s'agit de comparer des récoltes, des variétés, des parcours techniques, etc. Trois protocoles d'échantillonnage ont été testés comme préalables à la caractérisation du stade physiologique de lots d'abricots par des méthodes physiques destructives ou non. Le but recherché a été, hormis la réduction du nombre et du coût des analyses, l'obtention de lots homogènes représentatifs de la récolte et de son hétérogénéité, et la constitution de lots comparabl...

show abstract

“…S I is a MODL optimal discretization of S−S 1 . This interesting property is sufficient to adapt the dynamic programming algorithm presented in Fischer (1958), Lechevallier (1990), Fulton, Kasif, and Salzberg (1995) and Elomaa and Rousu (1996). We summarize in Table 1 this dynamic programming algorithm applied to the MODL discretization method.…”

Section: Optimal Algorithmmentioning

confidence: 99%

MODL: A Bayes optimal discretization method for continuous attributes

2006

View full text Add to dashboard Cite

While real data often comes in mixed format, discrete and continuous, many supervised induction algorithms require discrete data. Efficient discretization of continuous attributes is an important problem that has effects on speed, accuracy and understandability of the induction models. In this paper, we propose a new discretization method MODL 1 , founded on a Bayesian approach. We introduce a space of discretization models and a prior distribution defined on this model space. This results in the definition of a Bayes optimal evaluation criterion of discretizations. We then propose a new super-linear optimization algorithm that manages to find near-optimal discretizations. Extensive comparative experiments both on real and synthetic data demonstrate the high inductive performances obtained by the new discretization method.

show abstract

On Grouping for Maximum Homogeneity

Cited by 621 publications

References 3 publications

Optimization of difference pattern features in sub-arrayed monopulse antennas through and excitations matching procedure

Optimization of difference pattern features in sub-arrayed monopulse antennas through and excitations matching procedure

Méthodes d’échantillonnage de fruits appliquées à l’évaluation de la qualité de l’abricot

MODL: A Bayes optimal discretization method for continuous attributes

Contact Info

Product

Resources

About