There are several features that make Weibo a particularly interesting case for understanding the dynamics in social media as a platform for journalism’s work, especially in collective memory making and online resistance. First of all, the Weibo server is located in China and can be accessed by the majority of Chinese Internet users, which guarantees a large user basis. Due to the control of the Internet implemented by the Chinese government, accessing social media sites that are globally popular, such as Facebook, Twitter, and YouTube, is very difficult for the majority of Internet users in China, who do not have sufficient technological support to circumvent the Great Firewall. There are still a great number of people who do not even know the existence of such websites and services outside China. Taking Twitter as an example, those who tweet in Chinese are either technologically savvy or politically dissident, along with some overseas Chinese. Their voices are hardly heard by Chinese people who do not have access to Twitter and therefore have only a limited influence in China. People on Twitter have made attempts to post their tweets on China’s micro-blogging websites including Weibo, but due to strict censorship, these tweets and even user IDs are quickly detected and deleted by the web administrators. The accessibility of Weibo by the majority of Chinese people in China and other Chinese communities guarantees its influence on a remarkable amount of users. This is also the reason why the Chinese government is so anxious and uneasy about the fast growth of Weibo’s users and its influence.
ABSTRACT: In proposed a tensor factorization framework to simultaneously find the three sets of behavioral factors. Based on this framework, proposed system develops a numerical factorization model and another probabilistic factorization variant. Proposed work uses efficient algorithm based on large dataset and synthetic dataset for model propagation. This work enhances the existing work for blog propagation for user tweet with tweet dataset online. Online micro-blogging propagation model used stemming operation over twitter data to generate blog statement. Proposed works additionally implement tweet summarization for user view generation. Naïve bayes classifier differentiates the proposed work for blogging classification in the tweeter data.
Our inspiration comes from sold-start product recommendation using micro-blogging information: Wayne Xin Zhao and Ji-Rong Wen, They linked users across social networking sites and e-commerce websites for product recommendation and find a novel solution for cross-site cold-start product recommendation, which plans on propose results from e-commerce websites to users at social networking sites in “cold-start” circumstances. Our new scheme text mining algorithmic approach that utilized in proposed system that performed text mining over the users comments. This idea comes from Jayashri Khairnar, Mayura Kinikar proposed system which content text mining offers an approach to individuals and corporation information available on the internet analysis is a natural language processing and information extraction task that identifies the user’s perspectives explained in the form of positive, negative or impartial comments and quotes underlying the text .
We at that point build up a component based lattice factorization approach which can use the learnt client embeddings for frosty begin item suggestion. Test comes about on an expansive dataset developed from the biggest Chinese microblogging administration SINA WEIBO and the biggest Chinese B2C internet business site JINGDONG have demonstrated the viability of our proposed system.
Following the recent advances in both technology and social interaction, implementation of interactivity to large lecture rooms presents itself as a promising new method- ology to improve the learning and teaching process in academia. Namely, based on the underlying ideas of Web 2.0, learners should be able to collect and share online resources during a lecture, additional communication channels such as discussion forums, chat and micro- blogging helping to achieve interactivity on traditional face-to-face teaching. Building on such premises, first experiences have been acquired by the use of mobile devices and instant messaging in enhancing the learning and teaching behavior, with the help of a university wide available Learning Management System (LMS), which has accordingly been adapted and extended to the specific needs of supporting interactivity through mobile devices. The LMS is intended to use common and existing software and hardware (devices of the learners). The goal of the above research is to find out the potentials of interactivity in order to enhance students’ engagement in traditional face-to-face teaching in Higher Education. The paper describes the outcomes of the first experiments in implementing interactivity in Higher Education in such a framework within the Graz University of Technology (TU Graz) and comments on the methodology applied. The experiments, which have been performed during lec- turing within the course "Social Aspects of Information Technology" at the BSc level, attended by about 200 students, have shown that such kind of interactivity has a positive effect on the learners’ engagement.
In this paper, we have studied a novel problem, cross-site cold-start product recommendation, i.e., recommending products from e-commerce websites to microblogging users without historical purchase records. Our main idea is that on the e-commerce websites, users and products can be represented in the same latent feature space through feature learning with the recurrent neural networks. Using a set of linked users across both e-commerce websites and social networking sites as a bridge, we can learn feature mapping functions using a modified gradient boosting trees method, which maps users attributes extracted from social networking sites onto feature representations learned from e- commerce websites. The mapped user features can be effectively incorporated into a feature-based matrix factorization approach for cold start product recommendation.
As people are free to show their expression on anything using various microblogging sites like Twitter, Facebook, Instagram, Discussion forums and blogs. Mainly Microblogging and text messaging have emerged and become dominated tool over the web like whatsapp and others. Microblogging data is often used to share opinions and outlooks about the surrounding globe. The availability of social content generated on sites creates new opportunities to study public opinion and outlooks about the entity. This analysis we took anyone of the microblogging data for outlook classification and opinion building. The Outlook analysis is done on a microblogging data. The words are expressed in microblogging sites are compared with those in each other data that have been previously labeled as “positive”, or “negative”. After looking at these expressions, the algorithm then judges whether the text in the microblogging sites and is positive or negative based on the chances for each possibility. The overall objective of this paper is to determine the outlook or opinion of the microblogging text, whether it is positive or negative, which is extended to strength of polarity also this approach is used to obtain the considerable features and to analyzing the overall outlook for each object by computing the subjective standard for all the outlooks in the textual data.
Information-gathering has always been an important part to find what other person is thinking. Millions of users tweet on different aspects of life every day. Therefore microblogging websites are a very good source for polarity classification. We introduce a novel approach which automatically classifies the polarity of Twitter message. These messages so called tweets are classified as positive or negative or neutral. These results are useful for the customers or any general user who wants to research about the polarity of products before purchases, or it can be useful for the companies that want to analyze the reviews from people of their brands in the m arket. Most of the previous research on classifying the polarity of messages has tried to achieve some good results but have ignored the neutral tweets which lead to wrong polarity classification, so we have tried to solve this issue in our project. We present an approach for classifying the polarity of tweets using machine learning algorithms using a novel feature vector. Our training data cont ains publically available tweets which are obtained using twitter API‟s available. The following report shows the steps for preprocessing the dataset to achieve high accuracy. The novel feature vector of weighted unigrams that are used to train the machine learni ng classifiers is the main contribution of our project.
Embar et al.  define various functional and usability criteria that social influence scores should satisfy, and propose a multi-dimensional definition of social influence that satisfy these criteria. They consider many different dimensions of social influence: follower strength, activity, response rate or timeliness, etc. This work highlights the need to identify all-time influencers, which is a vivid instance of using the feature of time provided by micro-blogging services. We must note that social influence is a complex concept, which has various definitions and different evaluation metrics. Andrew McNeill and Pam Briggs  claim that social psychological theory can be used to in the qualitative analysis of Twitter data. They also point that when it comes to analysis the intensive social influence of tweets we should think about the how tweets can be influential by virtue of their content; the content includes emotion, themes, content category, and rhetorical strategy. Cha et al.  clearly present an in-depth description of two measures of a microblog’s social influence: repost and comment. These works are increasingly aware of the importance of a single microblog’s social influence.
In this paper, we have studied a novel problem, cross-site cold-start product recommendation, i.e., recommending products from e-commerce websites to microblogging users without historical purchase records. Our main idea is that on the e-commerce websites, users and products can be represented in the same latent feature space through feature learning with the recurrent neural networks. Using a set of linked users across both e- commerce websites and social networking sites as a bridge, we can learn feature mapping functions using a modified gradient boosting trees method, which maps users’ attributes extracted from social networking sites onto feature representations learned from e- commerce websites. The mapped user features can be effectively incorporated into a feature- based matrix factorization approach for cold start product recommendation. We have constructed a large dataset from WEIBO and JINGDONG. The results show that our proposed framework is indeed effective in addressing the cross-sitecold-start product recommendation problem. We believe that our study will have profound impact on both researchand industry communities. Currently, only a simple neutral network architecture has been employed for user and product embeddings learning. In the future, more advanced deep learning models such as
Brin& Page has introduced the Page Rank algorithm. Pre-computes a rank vector that provides a priori authority estimates for all of the nodes during a given graph. The node authority is independent of the attributes of every node associate degreed such an authority lives solely emerges from the topological structure of the graph. In particular, the authority of a node m depends on the number of incoming links and on the authority of the nodes that purpose to m with forward links. In this paper, a PageRank based model is planned to discover the most fashionable topics in micro-blogging supported users‟ interest relationship. The model first detects the favorite topics of every user with vote theory, then creates the links between topics with users‟ attentiveness relationship to create the „topic graph‟ in the entire micro-blogging social network, finally, ranks those topics with Page Rank algorithm to notice the foremost fashionable ones in micro-blogging.
In these days, product recommendation is a very important area to concentrates in increased sales for any ecommerce website. For example, Netflix has re- leased an interesting fact that about 75% of its subscriber’s watches are from recommendations system. There are many algorithms which focus on connecting the social media to ecommerce but none are focused on product recommendation by leveraging the social media information like demographic, micro-blogs, location, etc.
discover valuable client bunches by the clients' taking after shopping designs accepting that clients in a similar gathering offer comparable buy inclinations. Dormant gathering inclination, we regard a taking after client as a token and total every one of the followings of a client as an individual archive. Subsequently, we can remove idle client bunches having same interests (called "taking after themes"). d. Fleeting Attributes Temporal action examples are additionally used as they demonstrate the propensities and ways of life of the miniaturized scale blogging clients to some degree. There are a few relations between transient exercises examples and clients' buy inclinations. Worldly movement dispersions, we break down two sorts of transient action disseminations, every day and week after week action circulations. The day by day action circulation of a client is described by an appropriation of 24 proportions, and the emphasis shows the normal extent of tweets distributed inside the ith hour of a day by the client; likewise week by week movement conveyance of a client is portrayed by a dispersion of seven proportions, and the ith - proportion demonstrates the normal extent of tweets distributed inside the ith day of seven days by the client.
Abstract: Conceptual There is numerous essential applications for displaying of powerlessness and virality. To augment organization is reach comparably; popular client may employed by organizations to the promotion with viral substance or to proliferate positive substance about items. To direct crusading or to scatter lawmaker's messages generally they use on viral clients. We were geting a kick out of the chance to fuse even more fine-grained factors influencing the proliferation. At the point when open are confronting the social concentration, for mirroring the assessment of open fine-grained estimation is better. What's more, one may identify occasions by following those said by non-powerless clients and recognize bits of gossip in view of vulnerable client's connections with the substance. We rank clients by their virality (weakness) scores created by a virality demonstrate, select the best scored 1% clients as the anticipated viral (defenseless) clients, and signify the set by UPV (UPS). We adjust the V2s system by including one more feeling based factor of client conduct. This changed V2S joins the every single behavioural factor in one system, which mines the smaller scale blogging content, upgraded way. According to the outcome, investigation and execution result, our framework gives the 90 to 95 % of precise outcomes.
Twitter feed – this micro-blogging resource is used primarily to extend learners’ reading. I regularly ‘Tweet’ articles, sites and video content, streaming this into my blog so that my learners can have the opportunity to extend the sources and resources they use to learn. This has already had a massive impact upon the quality of classroom discussion (aided by increased further reading) and also many recent assignments. The use of the Twitter service for reading updates has had the greatest impact and uptake for the simplest and shortest amount of staff time.
The approach uses natural language processing techniques of Artificial Neural Network to extract features of interest from textual data retrieved from a microblogging platform in real-time and, hence, generate appropriate executable code for the Decision Science and get predetermined means of social communication. So by enriching semantic knowledge bases using Fuzzy Logic (for fitness approximation) for Opinion Mining in Big Data Applications with predetermined means, suggested user action decisions can be improved.
With the proliferation of micro-blogs, micro-blogging has been quickly gaining popularity and become an effective tool of com- munication for the quick organization of protests, help/advice, and sharing information from media sources, enabling unfamiliar groups of people to relay information of interest (Lee & Chan, 2012). This ability to disseminate information among social networks that lie out- side the control of institutions—such as the traditional media—has had a profound impact on traditional media’s agenda setting power immediately after an accident. The present study investigates the influence of micro-blogs on the major agenda-setting media in China during the immediate aftermath of a catastrophic railway accident. In particular, we investigate whether the singular agenda setting func- tion monopolized by the traditional Chinese media (e.g., Sun, 2010; Zhang & Zheng, 2012) is circumvented by micro-blogging following an accident that attracted national attention.