Real External Predictivity of QSAR Models. Part 2. New Intercomparable Thresholds for Different Validation Criteria and the Need for Scatter Plot Inspection

Chirico, Nicola; Gramatica, Paola

doi:10.1021/ci300084j

Cited by 431 publications

(276 citation statements)

References 21 publications

Supporting

Mentioning

268

Contrasting

Unclassified

Order By: Relevance

“…The performances in estimation of the top-three models with highest goodness-of-fit were assessed using the measures presented in Table 4 [23][24][25]. …”

Section: Assessment Of the Modelsmentioning

confidence: 99%

Szeged Matrix Property Indices as Descriptors to Characterize Fullerenes

Jäntschi

Bolboacă

2016

Ovidius University Annals of Chemistry

View full text Add to dashboard Cite

Abstract. Fullerenes are class of allotropes of carbon organized as closed cages or tubes of carbon atoms. The fullerenes with small number of atoms were not frequently investigated. This paper presents a detailed treatment of total strain energy as function of structural feature extracted from isomers of C40 fullerene using Szeged Matrix Property Indices (SMPI). The paper has a two-fold structure. First, the total strain energy of C40 fullerene isomers (40 structures) was linked with SMPI descriptors under two scenarios, one which incorporate just the SMPI descriptors and the other one which contains also five calculated properties (dipole moment, scf-binding-energy, scf-core-energy, scf-electronic-energy, and heat of formation). Second, the performing models identified on C40 fullerene family or the descriptors of these models were used to predict the total strain energy on C42 fullerene isomers. The obtained results show that the inclusion of properties in the pool of descriptors led to the reduction of accurate linear models. One property, namely scf-binding-energy proved a significant contribution to total strain energy of C40 fullerene isomers. However, the top-three most performing models contain just SMPI descriptors. A model with four descriptors proved most accurate model and show fair abilities in prediction of the same property on C42 fullerene isomers when the approach considered the descriptors identified on C40 as the predicting descriptors for C42 fullerene isomers.

show abstract

“…The performances in estimation of the top-three models with highest goodness-of-fit were assessed using the measures presented in Table 4 [23][24][25]. …”

Section: Assessment Of the Modelsmentioning

confidence: 99%

Szeged Matrix Property Indices as Descriptors to Characterize Fullerenes

Jäntschi

Bolboacă

2016

Ovidius University Annals of Chemistry

View full text Add to dashboard Cite

show abstract

“…The comparable thresholds used in this study for different validation criteria have been rigorously previously determined [25,28] . Other statistical parameters [29] were used for the external test set: (i) squared correlation coefficient (r …”

Section: Model Validationmentioning

confidence: 99%

Structure-flammability relationship study of phosphoester dimers by MLR and PLS

2016

View full text Add to dashboard Cite

SbstractPolyphosphonates and polyphosphates having good flame retardancy represent an important class of organophosphorus based polymer additives. In this analysis the flammability of 28 previously synthesized polyphosphoesters, modelled as dimmers, was explored using the multiple linear regression (MLR) and Partial Least Square (PLS) methodology. The statistical quality of the final MLR and PLS models was estimated using the following parameters: the squared correlation coefficient (r 2 training = 0.917 and 0.976), the training root-mean-square errors (RMSE tr = 0.029 and 0.016) and the leave-seven-out cross-validation correlation coefficient (q 2 L70 = 0.748 and 0.881), respectively. External validation was checked for a test set of seven compounds using several criteria. The MLR models had somewhat inferior fitting results. The final MLR and PLS models can be used for the estimation of limiting oxygen index (LOI) values of new polyphosphoester structures. The presence of phosphonate groups and increasing molecular branching in an isomeric series favour the dimer flammability.

show abstract

“…To test the robustness and predictive power of the model the concordance correlation coefficient (CCC) [15] (having the thresholds values higher than 0.85, as they have been rigorously determined by a simulation study [16]). …”

Section: Model Validationmentioning

confidence: 99%