An Efficient Two-Layer Framework for Tour Sense Recommendation

Abstract

The paper shows a cutting edge prototype system, which can recommend most comprehensive travel plans that include brand new, diverse latest interest factors (POIs). It systematically gathers and analyzes data on thousands of cutting-edge tourism destinations and geographical nodes. Tour feel is a recommendation framework which examines the preference information modern day diverse tourists based totally on the transport records collected from various towns. Humans can get properly-in shape path plans, which consist of a sequence brand new points cutting-edge interest (POIs) primarily based on vacationers' constraints and goals, way to the advancement ultra-modern excursion recommendation. Multiple scenic spots and entrances are common in big-scale POIs, which are known as brilliant-POIs in this article. Maximum contemporary excursion advice algorithms, alternatively, forget the big expertise determined inside super POIs. A layer machine, which takes into account the ultra-modern route architecture (outer model) and panoramic routes within outstanding POIs (internal model). A full-on-integration-based Embedded Keep VND algorithm is used to merge templates. A greedy randomized adjustable path advent system (grasp) for local improvements in the internal version of the outer version and variable neighborhood descent (VND). The first rate-POI is considered "meta node" in the development of trendy outer routes. Dijkstra Algorithm with Pruning dynamically reviews the most satisfying path within the exquisite POI in the internal model to conform to the outer direction. The obtained end results provide the traveller with the ultimate quality route and the focus point chosen by the users.

Design of University Archives Business Data Push System Based on Big Data Mining Technology

“…Since the university archives business data push system is aimed at users in different geographical locations, the business publisher on the PC side also publishes business in different work areas [10], [11]. The work nodes are distributed in different locations, so the university archives business data push system designed by the proposed method adopts a star topology structure, as shown in Fig.…”

“…2) Ranking of recommended items: The ranking of recommended items comprehensively considers factors such as feature similarity between users, behavior sequence similarity, popularity of business data and timeliness of business data, calculates the ranking weight of business data, sorts it, and finally generates list of recommended items of file business data. The specific process is divided into the following steps: a) Calculate the total feature similarity of users associated with data: One file may be related to multiple users, so the overall feature similarity of file business data is the user behavior sequence similarity between the generated characteristics of the current user and those of all neighboring users containing the data which is calculated as follows: (11) b) Calculate the total similarity of the behavior sequence associated with the data: One file business may be related to multiple user behavior sequences, so the similarity of one file business behavior sequence is the behavior sequence and prediction sequence associated with the similarity which is calculated as follows: ∑…”