The combined analysis of polarized DIS and SIDIS data is performed in NLO QCD. The new parametrization on polarized PDFs is constructed. The uncertainties on PDFs and their first moments are estimated applying the modified Hessian method. The especial attention is paid to the impact of novel SIDIS data on the polarized distributions of light sea and strange quarks. In particular, the important question of polarized sea symmetry is studied in comparison with the latest results on this subject. : 13.65.Ni, 13.60.Hb, 13.88.+e Since the observation of the famous spin crisis in 1987 [1] one of the most intriguing and still unsolved problems of the modern high energy physics is the nucleon spin puzzle. The key component of this problem, which attracted the great both theoretical and experimental efforts during many years is the finding of the polarized parton distributions functions (PDFs) in nucleon.
PACSThe analysis of data on inclusive polarized DIS enables us to extract such important quantities as the singlet ∆Σ(x, Q 2 ) and nonsinglet ∆q 3 (x, Q 2 ), ∆q 8 (x, Q 2 ) combinations of the polarized PDFs, and, thereby, the sums of valence and sea PDFs ∆q + ∆q ≡ ∆q V + 2∆q. Besides, dealing with DIS data, the gluon helicity distribution ∆G(x, Q 2 ) is determined due to the evolution in singlet sector and weak dependence on ∆G of the polarized structure function g 1 in NLO QCD. However, even for singlet combinations ∆u + ∆ū, ∆d + ∆d, ∆s + ∆s, considered as well-determined within DIS, we still meet the problem: there are only two equations corresponding to inclusive asymmetries A 1 measured on proton and deuteron targets from which we should determine three unknown combinations ∆q 3 (x), ∆q 8 (x) and ∆Σ(x) (or, alternatively, ∆u + ∆ū, ∆d + ∆d, ∆s + ∆s). So, it is unavoidable to involve some additional assumptions performing the fitting procedure for the purely inclusive DIS data. Moreover, DIS data can not help us to solve the important problem of valence and sea PDFs separation.The basic 4 process which enables us to solve these problems is the process of semiinclusive DIS (SIDIS). However, until recently the quality of the polarized SIDIS data was rather poor, so that its inclusion in the analysis did not helped us [2] to solve the main task of SIDIS measurements: to extract the polarized sea and valence PDFs of all active flavors. Only in 2004 the first polarized SIDIS data with the identification of produced 1 email: sisakian@jinr.ru 2 email: shev@mail.cern.ch 3 email: ivon@jinr.ru 4 As long as no neutrino factory is built and no hyperdense polarized target is created, we cannot study the DIS processes with the neutrino beam, which would enable us to find the polarized valence and sea quark distributions separately.