This paper proposes a new paradigm for sentiment analysis: translation from text documents to a set of sentiment units. The techniques of deep language analysis for machine translation are applicable also to this kind of text mining task. We developed a high-precision sentiment analysis system at a low development cost, by making use of an existing transfer-based machine translation engine.
The effects of neutron and ion irradiations on deuterium (D) retention in tungsten (W) were investigated. Specimens of pure W were irradiated with neutrons to 0.3 dpa at around 323 K and then exposed to high-flux D plasma at 473 and 773 K. The concentration of D significantly increased by neutron irradiation and reached 0.8 at% at 473K and 0.4 at% at 773 K. Annealing tests for the specimens irradiated with 20 MeV W ions showed that the defects which play a dominant role in the trapping at high temperature were stable at least up to 973 K, while the density decreased at temperatures equal to or above 1123 K. These observations of the thermal stability of traps and the activation energy for D detrapping examined in a previous study (≈1.8 eV) indicated that the defects which contribute predominantly to trapping at 773 K were small voids. The higher concentration of trapped D at 473 K was explained by additional contributions of weaker traps. The release of trapped D was clearly enhanced by the exposure to atomic hydrogen at 473 K, though higher temperatures are more effective for using this effect for tritium removal in fusion reactors.2
We have studied the dielectronic recombination process in He-like Fe ions and have obtained the resonant strengths of the KLn (3 n 5) resonances. This measurement was performed with the use of an electron beam ion trap by measuring the x-ray energy emitted from highly charged ions simultaneously with the electron beam energy scanned during the measurement. The total resonant strengths obtained are 5.0×10 −19 , 2.1×10 −19 and 1.1×10 −19 cm 2 eV, for KLM, KLN and KLO, respectively.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.