Classification models, predictive applications

The KNN method is probably the simplest classification method to understand. It is most commonly applied to a principal component space. In this case, calibration is achieved by simply constructing a PCA model using the calibration data, and choosing the optimal number of PCs (A) to use in the model. Prediction of an unknown sample is then done by calculating the PC scores for that sample (Equation 8.57), followed by application of the classification rule. 

To increase the predictivity of decision tree classification models, statistical tools such as boosting [62] have been employed in the context of decision tree classification. The application of this technique in predicting structure-property relationships showed to significantly increase the accuracy and robustness of the obtained decision tree models however, this is at the cost of comprehensiveness of the model and the computational speed of model generation [56]. 

In Fig. 7 we have shown visually the best performances of multiclass OVR SVM, VVRKFA and DAG NPPC models obtained experimentally in the 10-fold CV method. The Fig. 7 shows that the OVR SVM and VVRKFA models with RBF kernels achieve high classification accuracy in all the 10 individual folds. In view of all these results, it is observed that the model trained with OVR multiclass SVM and VVRKFA with Gaussian RBF kernel may be suitable for black tea quality prediction application. 

A key question is as follows Can SE and DSE, as an information theoretic approach to descriptor comparison and selection, be applied to accurately classify compoimds or to model physiochemical properties To answer this question, two conceptually different applications of SE and DSE analysis will be discussed here and related to other studies. The first application explores systematic differences between compound sets from synthetic and natural sources." The second addresses the problem of rational descriptor selection to predict the aqueous solubility of synthetic compounds." For these purposes, SE or DSE analysis were carried out, and in both cases, selected descriptors were used to build binary QSAR-like classification models. 

The possibilities for the application for neural networks in chemistry arc huge [10. They can be used for various tasks for the classification of structures or reactions, for establishing spcctra-strncturc correlations, for modeling and predicting biological activities, or to map the electrostatic potential on molecular surfaces. 

Exploratory data analysis shows the aptitude of an ensemble of chemical sensors to be utilized for a given application, leaving to the supervised classification the task of building a model to be used to predict the class membership of unknown samples. 

MacKay s textbook [114] offers not only a comprehensive coverage of Shannon s theory of information but also probabilistic data modeling and the mathematical theory of neural networks. Artificial NN can be applied when problems appear with processing and analyzing the data, with their prediction and classification (data mining). The wide range of applications of NN also comprises optimization issues. The information-theoretic capabilities of some neural network algorithms are examined and neural networks are motivated as statistical models [114]. 

The applicability of PBPK models can be described in the context of the BDDCS classification [24]. PBPK models are very predictive for class 1 and class 2 compounds. However for poorly soluble compounds, the use of aqueous solubility is shown to be inadequate for reliable prediction of oral absorption in physiologically based models [7]. In such cases, it is recommended to use solubility measured in simulated intestinal fluids (FeSSIF, FaSSIF). Such data proved to be very relevant to simulate the oral absorption of BCS 2 (low solubility, high permeability) compounds [25]. 

Cross-validation is an alternative to the split-sample method of estimating prediction accuracy (5). Molinaro et al. describe and evaluate many variants of cross-validation and bootstrap re-sampling for classification problems where the number of candidate predictors vastly exceeds the number of cases (13). The cross-validated prediction error is an estimate of the prediction error associated with application of the algorithm for model building to the entire dataset. 

Simplistic models with few rules have proven to be highly beneficial in predicting basic pharmacokinetic properties without extensive computational analysis. The Lipinkski rule of 5 for predicting which compounds are likely to show good or poor absorption properties is one example that has found wide application in industry and there are others. Egan et al. (2000) reported that compounds possessing a PSA of <148.1A2 and log Kow above 5.88 would be poorly absorbed. Veber et al. (2002) proposed a simple classification system for oral bioavailability, where compounds... 

---