Design And Implementation Of Personalized Recommendation Engine Based On Improved Trie Tree And AC Algorithm In Android Platform

With the rapid development of science and technology, the computer havedeveloped to the Pocket PC from the initial monster, volume reduction has notweakened its ability, but the function is more complete and faster. With theimprovement of function of PPC, more people depend on the handheld computer dealtheir personal things, the user has to carry the cell phone and PPC at all times. Thus,the smart manufacturers put the Pocket PC operating system ported to mobile phones,so the smart phone was born. It is like a personal computer, it has an independentoperating system that allows users to install their own office software, games, andother third service provider programs. These applications continuous expand thephone’s feature. Also, it can access the internet by WIFI. Once accessing the network,users will be able to acquire any desired information by search engine. In the searchfield, the outcome has been very obvious. In foreign countries, Google be way ahead.In China, Baidu outshines. However, in the recommendation domain, at least so far,there is no recommendation engine is fully deserved the market leader.Currently, Android and IOS are the main smart phone operating system on themarket. According to the latest data show that as of November2012, the Androidsystem has occupied76%of the global smart phone operating system market, andmarket share in China has reached90%.This topic is based on the above environment, try to develop a recommendationengine that based on android platform. The analysis engine mainly uses improved Trietree and AC algorithms to implement.Trie tree, also known as dictionary, word search tree or trie tree, is a multi-treestructure for fast lookup. Trie tree mainly contains three basic characteristics:First, theroot node does not contain characters, the other nodes only contain one character;Second, from the root to the specified node, stitching all the characters which is on thepath together then represents a string; Third, all child nodes of each node contained in the string must be different. The core idea of Trie tree is exchanging space for time,using common prefix string to reduce query time to improve efficiency.Its typicalapplication is counting and sorting the large numbers of strings (but not limited tostring), Therefore, search engines are often used for text word frequency statistics.AC algorithm called Aho–Corasick string matching algorithm, was invented byAlfred V. Aho and Margaret J. Corasick. The so-called string matching is to find outthe words that is matching the given pattern string from a text. AC algorithm is aclassic multi-pattern matching algorithm that can locate all the elements in the text ona limited set of pattern strings (the dictionary), and matching all modes at the sametime. The key to the AC algorithm is constructing the failed node based on the Trietree. Each node has a failure node, when looking to the current node does not match,then jump directly to the failure node of the current node continue searching.The research of personalized recommendation engine is based on Androidmobile phone.Firstly, It gets the user’s message content (only English text) on themobile phone and analyze the content and get the result topic,then push information touser through the result topic. The biggest difference of the recommendation engineand other recommendation engine is that It analyzes the existing data on the phone,rather than that of the browser-based. The recommendation engine of this topicachieves more efficient use of the user Information. It makes the suggestion featureextend from the site to the mobile and allows the user to get the desired informationwhat you always want in the usual routine. The recommendation engine combinedwith the Android mobile phone platform provides a broad development space for itsfuture application prospect.