Atomic parameter-pattern recognition method

When an atom is in isolated state, the energy levels of its electrons are the characteristic values for this atom. When this atom combines with other atoms to form molecules or condensed states, the electronic energy levels will be changed. And that of valence electrons changes more evidently. For this reason, we cannot use the data of ionization potential directly for the computation of the physico-chemical properties of chemical substances. But the values of the ionization potentials of isolated atoms should be eventually correlated to the energy levels in molecules or condensed states in some complicated manner. So the values of ionization potential of isolated atoms of elements should also be a useful atomic parameter in atomic parameter pattern recognition method. 

Experimental eonfirmation or negation of the suspicious phase diagrams found by atomic parameter-pattern recognition method If some published phase diagrams have been found to be an outlier or suspicious, we should do experimental work to confirm or negate our suspicion. 

Atomic Parameter-Pattern Recognition Method for Phase Diagram Prediction... 

The formation of intermediate compounds is one of the most important factors affecting the geometry of phase diagrams. Atomic parameter-pattern recognition method is an effective technique dealing with the assessment or prediction problems about the formability of intermediate compounds. 

Atomic parameter-pattern recognition method is rather effective for the computerized prediction of the crystal types of unknown intermediate phases. 

Methods can be based on some preconceived concept of bonding, with ionic and covalent extremes, or on pattern recognition based on the periodic table. Miscellaneous methods of limited applicability link bond strength with other physical properties. The a priori calculation of heats of formation by wave mechanics is possible in theory. In practice, the most widely applied method incorporates experimental data to derive atom or bond parameters which can then be used for calculations on closely related compounds. 

In recent years, we have used this comprehensive strategy to make the computerized assessment of a series of phase diagrams. By this strategy, we have found that the coordination of atomic parameter method and various pattern recognition techniques (including SVM) is rather effective to find out the outliers or suspicious phase diagrams [30 23 26 32],... 

---