Design And Implemention Of Customer Advertising Optimizer

With the increasing popularity of the Internet, the size of Internet advertising increases,in 2015, China’s online advertising market has reached 209.37 billion yuan.of which the proportion of search engine advertising has been occupying the first place.In the process of advertising,due to the limitation of advertisers’own condition,they have problems of getting advertising data,analysis of advertising data and getting optimization suggestions for advertising. To solve these problems, the author’s company decided to develop a customer advertising optimization tool on the basis of the existing advertising platform, providing the advertising data and analysis services based on the current situation,and providing advertising optimization suggestions for advertisers.Customer advertising optimizer provides visual data display and optimization suggestions,can provide comprehensive data and accurate optimization suggestions. The tool contains pull-data module, multi-layer cache, monitoring logs,data display module, adjust price module, keywords recommendation module.For many kinds of data,big data volume, complex data structures, uses Nosql (MongoDB) database for data storage. For users access to the tool frequently and the capacity of MongoDB is limited, the system design a multi-layer cache architecture,using Memcached caching data that has been requested in MongoDB to prevent system from performance consumption, improve the access speed of system.In the keyword recommendention module, the author uses LFM (latent factor model) to complete the relevant algorithm design and development work.In the process of constructing the customer advertising optimizer. First of all, the author participated in the overall design of the customer advertising optimiser and prototype design of keyword recommendention module. Secondly, the author independently completed the customer advertising optimiser requirements’gathering and analysis. In the phase of detailed design and implementation, the author independently completed pull-data module,monitoring logs module,data display module,adjust price module,keyword recommendention module. In the test phase, the author completed the unit test of all modules.At present, the customer advertising optimiser has been successful on-line. These modules completed by the author are running normally, and the function of which is stable. All jobs of the author have met the expected goal.