Abstract: Multiple sources of reviews for the same product or service (e.g. hotels, restaurants, clinics, hair saloon, etc.) are utilized to provide a trustworthiness score. Such a score can clearly identify hotels with evidence of review manipulation, omission and fakery and provide the user with a comprehensive understanding of the reviews of a product or establishment. Three types of information are used in computing the score: spatial, temporal and network or graph-based. The information is blended to produce a representative set of features that can reliably produce the trustworthiness score. The invention is self-adapting to new reviews and sites. The invention also includes a validation mechanism by crowd-sourcing and fake review generation to ensure reliability and trustworthiness of the scoring.
Type:
Grant
Filed:
September 9, 2015
Date of Patent:
October 2, 2018
Assignee:
STC.UNM
Inventors:
Shuang Luan, Abdullah Mueen, Michalis Faloutsos, Amanda J. Minnich
Abstract: A backup computing device detects an interruption while receiving a backup copy of a transaction log of a primary database and directs a secondary computing device to continue generation of the backup copy of the transaction log, based on the copy of the transaction log of the primary database. A primary computing device directs a secondary computing device to generate a backup copy of the data file and the primary computing device generates a backup copy of the transaction log, of a primary database. A primary computing device accesses a snapshot of a plurality of snapshots of a primary database, and generates a first portion of a backup copy of the database. The primary computing device directs a second computing device to generate a non-overlapping portion of the backup copy of the database based on a second snapshot of the plurality of snapshots.
Type:
Grant
Filed:
December 21, 2017
Date of Patent:
September 18, 2018
Assignee:
International Business Machines Corporation
Abstract: Novel methods are described herein to provide an OLAP database system that performs lock-less bulk insertion while maintaining high query performance and minimizing additional storage requirements. Aspects of the claimed subject matter include novel methods for merging two index structures while maintaining query consistency without the use of a versioning scheme by using copies of vertical segments of an index tree page; a method to prepare and merge two index structures that operates within storage constraints; and a method to execute queries concurrently while index maintenance is in progress while still producing consistent results. According to an aspect of the present disclosure, a flexible index merge mechanism is provided that merges data from incremental indices to a primary data index in three phases: a load phase, a copy phase, and a replace phase.