Spectral classification of uranium i energy levels using pattern-recognition techniques

Previously unknown uranium i energy levels have been classified according to configuration using pattern-recognition techniques. Four features, the energy level, Landé g factor, quantum number J, and ${}_{}{}^{235}{}_{}{}^{}\mathrm{U}$ isotope shift, have been used to describe each level. Additionally, a fifth feature, the ${}_{}{}^{234}{}_{}{}^{}\mathrm{U}$ isotope shift, has been employed to describe most levels. Thirty-seven levels have been assigned with high certainty. In addition, four-feature classification versus five-feature classification is discussed.