Project topics and materials
Get Complete Project Material File(s) Now! »
Predicting Trend Makers and Spotters
By considering four types of features, we have been able to find statistically significant similarities and differences among trend spotters, trend makers, and typical users. Now we study to which extent these features are potential predictors of whether users are trend spotters (makers), and do so in two steps:
1. We model trend spotter (maker) score as a linear combination of the features.
2. We predict trend spotter (maker) using a logistic regression and a machine learning model: Support Vector Machines (SVM).
Upon the set of 140 trend makers, 671 trend spotters and 1,705 typical users (identified in the previous section), we now run our predictions.
Regression Models
Before running the regression, we compute the (Pearson) correlation coefficients between each pair of predictors (Table 3.3). As one expects, we find that different types of activities are correlated (i.e., high positive correlation between the number of followees, daily uploads, daily votes, and content diversity). Attracting followers is correlated more with uploading content (i.e., positive correlation between the number of followers and daily uploads) rather than voting content (i.e., no significant high correlation between the number of followers and daily votes).
Support Vector Machines (SVM)
We formulate the task of predicting trend spotters (makers) as a binary classification problem, where the response variable is whether a user’s trend spotter (maker) score is greater than or equal to zero. To our sample of 671 trend spotters and 140 trend makers, we add an equal number of typical users (those 1,705 users have been identified in the previous section). By construction, the resulting sample is balanced (the response variable is split 50-50), and interpreting the results becomes now easy as the accuracy of a random prediction model would be 50%.
We split randomly each set of samples into two subsets, 80% of them are used for training and 20% for testing. We apply SVM 3 on the input of the same seven features previously used in the regressions to predict trend spotter and trend maker scores. We compare the results with those obtained by the previous logistic regression model, and we show their prediction performance by ROC (Receiver Operating Characteristic) curve (Figure A.1), AUC (area under the ROC curve), and accuracy (Table A.5). In a ROC curve plot, an ideal prediction model is expected to achieve a high true positive rate but with a low false positive rate. In Figure A.1, the point (0,1) corresponds to the perfect prediction while (1,0) corresponds to the worst, and the diagonal line reflects the baseline of a random guess. Points above the baseline indicate good prediction results [71].
Practical Implications
The ability of identifying trend spotters and trend makers has implications in designing recommender systems, new products and user interfaces.
Recommender Systems. Every user has different interests and tastes and, as such, might well benefit from personalized suggestions of content. These suggestions are automatically produced by so-called “recommender systems”. Typically, these systems produce recommendations people might like by equally weighting all user ratings. Given that trend spotters are effective social filters, one could imagine to weight their ratings more those from typical users to construct a new recommender system. New Products. Some web services (e.g., 99designs4) provide a platform to crowd source design work, where clients submit their requests and designers try to fulfill them. Since trend spotters and trend makers are “fashion leaders”, soliciting their early feedbacks might result into avoiding mistakes when designing new products. Often, at design stage, costs of correcting minor mistakes are negligible, while, at production stage, they become prohibitive.
User Interfaces. Trend spotters and trend makers do not connect to as many users as one would expect. That is likely because it is hard for iCoolhunt users to be aware of what others are up to. The user interface does not come with clear-cut “social features” that create a sense of connection and awareness among users as much as Facebook or Twitter sharing features do (as we have detailed in the Application section).
Table of contents :
List of Figures
List of Tables
1 Introduction
1.1 Motivations
1.2 Research Problems and Contributions
1.3 Organization of the thesis
2 Background and RelatedWork
2.1 Trends in Social Media
2.1.1 Identification
2.1.2 Characterization
2.1.3 Influentials
2.2 Recommender Systems
2.2.1 User Preferences
2.2.2 Collaborative Filtering
2.2.3 Evaluation
2.3 Summary
3 Who Creates Trends
3.1 Background on the Mobile Social Application
3.2 Dataset
3.2.1 Uploads and Votes
3.2.2 Geography
3.2.3 Following
3.3 Identifying Trend Makers and Spotters
3.3.1 Defining Trend Spotters and Trend Makers
3.3.2 Characterizing Trend Spotters and Trend Makers
3.3.3 Who trend spotters and trend makers are
3.4 Predicting Trend Makers and Spotters
3.4.1 Regression Models
3.4.2 Support Vector Machines (SVM)
3.5 Discussion
3.5.1 Theoretical Implications
3.5.2 Practical Implications
3.6 Summary
4 Personalizing Trends
4.1 Background
4.2 Trend-aware Recommendation
4.2.1 Identify trend makers and trend spotters
4.2.2 Identify Trends
4.2.3 Recommend Trends
4.3 Evaluation
4.3.1 Classifying users into trend spotter(maker) classes
4.3.2 Determining whether an item is a trend or not
4.3.3 Recommending trends
4.4 Discussion
4.5 Summary
5 Serendipitous Recommendations
5.1 Background
5.2 Our Proposals
5.2.1 Basic Algorithms
5.2.2 Beyond User History
5.2.3 Beyond User Routine
5.3 Evaluation
5.3.1 Evaluation Metrics
5.3.2 Data
5.3.3 Validating Modeling Assumptions
5.3.4 Accuracy vs. Serendipity
5.4 Discussion
5.5 Summary
6 Conclusion and Future Work
6.1 Thesis Contributions
6.2 Future Work
Bibliography
