Search

Product reviews are of great commercial value for online shopping market. The identification of customer opinions from product reviews is helpful to improve the marketing decisions of customers, sellers and producers. This paper proposes a novel framework for summarizing customer opinions from product reviews. Firstly, our framework identifies grammatically and semantically meaningful phrases which contain product attributes and their corresponding opinions from original product reviews by using grammar rules and the latent Dirichlet allocation (LDA) model. Secondly, our framework generates readable and simple summaries from the identified phrases automatically by using the deep neural network. The summaries provide users the valuable opinions on product attributes. Moreover, our framework provides an interactive mode for users to choose product attributes which they are interested for generate personalized summaries to help users focus on the most concerned opinions. Experimental results on six datasets demonstrate effectiveness of our framework. [ABSTRACT FROM AUTHOR]

Asness and Brown acknowledged that having a stateful model might improve the decision-making power of the AB model; specifically, that knowing if the puck is in the offensive zone, neutral zone, or defensive zone would vary their recommendation. When one team has pulled the goalie, the scoring rate becomes 1.97% per 10-second interval for the team with the sixth attacker, and 4.3% for the team that keeps its goalie. If the time is between 330 seconds and less than or equals to 550 seconds, the coach should pull the goalie when the puck is in the offensive zone, and put the goalie back when the puck is in the neutral or defensive zone. [Extracted from the article]

We present a new method to measure neutrino masses using the dark matter-neutrino relative velocity. The relative motion between dark matter and neutrinos results in a parity-odd bispectrum which can be measured from cross-correlation of different cosmic fields. This new method is not affected by most systematics which are parity even and not limited by the knowledge of optical depth to the cosmic microwave background. We estimate the detectability of the relative velocity effect and find that the minimal sum of neutrino masses could be detected at high significance with upcoming surveys. [ABSTRACT FROM AUTHOR]

We discuss the possibility of studying diffuse baryon distributions with the kinematic Sunyaev-Zel'dovich (kSZ) effect by correlating cosmic microwave background (CMB) temperature fluctuations with density fluctuations from 21 cm intensity mapping (IM). The biggest challenge for the cross-correlation is the loss of large-scale information in IM, due to foregrounds and the zero spacing problem of interferometers. We apply the 3D tidal reconstruction algorithm to restore the lost large-scale modes, which increases the correlation by more than a factor of 3. With the predicted foreground level, we expect a ~20σ detection of kSZ signal for 0.8≲z≲2.5 with CHIME and Planck, and a ~40σ detection with HIRAX and Planck. The significance can be greatly increased with next-generation facilities of higher spatial resolutions. [ABSTRACT FROM AUTHOR]

21 cm intensity mapping has emerged as a promising technique to map the large-scale structure of the Universe, at redshifts z from 1 to 10. Unfortunately, many of the key cross-correlations with the cosmic microwave background and photo-z galaxies have been thought to be impossible due to the foreground contamination for radial modes with small wave numbers. In this paper, we apply tidal reconstruction to the simulated 21 cm fields and recover the lost large-scale radial modes successfully. We estimate the detectability of the cross-correlation signals and find they can be detected at high significance with current 21 cm experiments. The tidal field reconstruction method opens up a new set of possibilities to probe the Universe and is extremely valuable not only for 21 cm surveys but also for cosmic microwave background and photometric-redshift observations. [ABSTRACT FROM AUTHOR]

We apply nonlinear reconstruction to the dark matter density field in redshift space and solve for the nonlinear mapping from the initial Lagrangian position to the final redshift space position. The reconstructed anisotropic field inferred from the nonlinear displacement correlates with the linear initial conditions to much smaller scales than the redshift space density field. The number of linear modes in the density field is improved by a factor of 30-40 after reconstruction. We thus expect this reconstruction approach to substantially expand the cosmological information including baryon acoustic oscillations and redshift space distortions for dense low-redshift large scale structure surveys including for example SDSS main sample, DESI BGS, and 21 cm intensity mapping surveys. [ABSTRACT FROM AUTHOR]

We present a direct approach to nonparametrically reconstruct the linear density field from an observed nonlinear map. We solve for the unique displacement potential consistent with the nonlinear density and positive definite coordinate transformation using a multigrid algorithm. We show that we recover the linear initial conditions up to the nonlinear scale (rδrδL>0.5 for k≲1 h/Mpc) with minimal computational cost. This reconstruction approach generalizes the linear displacement theory to fully nonlinear fields, potentially substantially expanding the baryon acoustic oscillations and redshift space distortions information content of dense large scale structure surveys, including for example SDSS main sample and 21 cm intensity mapping initiatives. [ABSTRACT FROM AUTHOR]

Measurements of neutrino mass in cosmological observations rely on two-point statistics that are hindered by significant degeneracies with the optical depth and galaxy bias. The relative velocity effect between cold dark matter and neutrinos induces a large scale dipole in the matter density field and may be able to provide orthogonal constraints to standard techniques. We numerically investigate this dipole in the TianNu simulation, which contains cold dark matter and 50 meV neutrinos. We first compute the dipole using a new linear response technique where we treat the displacement caused by the relative velocity as a phase in Fourier space and then integrate the matter power spectrum over redshift. Then, we compute the dipole numerically in real space using the simulation density and velocity fields. We find excellent agreement between the linear response and N-body methods. Utilizing the dipole as an observational tool requires two tracers of the matter distribution that are differently biased with respect to the neutrino density. [ABSTRACT FROM AUTHOR]

We study the nonlinear E-mode clustering in Lagrangian space by using large scale structure N-body simulations and use the displacement field information in Lagrangian space to recover the primordial linear density field. We find that, compared to Eulerian nonlinear density fields, the E-mode displacement fields in Lagrangian space improves the cross-correlation scale k with initial density field by a factor of 6-7, containing 2 orders of magnitude more primordial information. This illustrates ability of potential density reconstruction algorithms, to improve the baryonic acoustic oscillation measurements from current and future large scale structure surveys. [ABSTRACT FROM AUTHOR]

Galaxy-galaxy strong gravitational lenses have become a popular probe of dark matter (DM) by providing a window into structure formation on the smallest scales. In particular, the convergence power spectrum of subhalos within lensing galaxies has been suggested as a promising observable to study DM. However, the distances involved in strong-lensing systems are vast, and we expect the relevant volume to contain line-of-sight (LOS) halos that are not associated with the main lens. We develop a formalism to calculate the effect of LOS halos as an effective convergence power spectrum. The multilens plane equation couples the angular deflections of consecutive lens planes, but by assuming that the perturbations due to the LOS halos are small, we show that they can be projected onto the main-lens plane as effective subhalos. We test our formalism by simulating lensing systems using the full multiplane lens equation and find excellent agreement. We show how the relative contribution of LOS halos and subhalos depends on the source and lens redshift, as well as the assumed halo and subhalo mass functions. For a fiducial system with fraction of DM halo mass in substructure fsub=0.4% for subhalo masses [105-108]M⊙, the interloper contribution to the power spectrum is at least several times greater than that of subhalos for source redshifts zs≳0.5. Furthermore, it is likely that for the SLACS and BELLS lenses the interloper contribution dominates: fsub≳2% (4%) is needed for subhalos to dominate in SLACS (BELLS), which is higher than current upper bounds on fsub for our mass range. Since the halo mass function is better understood from first principles, the dominance of interlopers in galaxy-galaxy lenses with high-quality imaging can be seen as a significant advantage when translating this observable into a constraint on DM. [ABSTRACT FROM AUTHOR]