Journal Press India^®

Author Details ( * ) denotes Corresponding author

Hashtag investor is a system that can analyze twitter data to generate useful information including some predictions. Machine learning techniques have been used for this research which falls into data mining to archive sentiment analysis to categorize and identify tweets based on the contents. Twitter has an enormous collection of data. If these data is converted into some useful information, accurate decisions can be made using this data. That is our main objective, which can be very helpful to users, and this system works with respect to four specific objectives. One objective is sentimental analysis of twitter data and finding false tweets. Supervised learning has been used and NLTK and also the naïve Bayes classifier has been used as techniques. The output will be display percentage wise, negative positive and neutral percentages of the given keyword. Twitter data is analyzed according to the given keyword. False tweets identification is done by analyzing user profile. If the user profile criteria does not match with our assumptions this profile is marked as a fake profile. Second objective is comparing two similar products and getting the popularity according to the time. The output is displayed by charts. Similar keywords will be grouped. Clustering algorithms has been used for grouping. Our forth objective is finding some latest ongoing events and the number of users who were active at certain time periods, ARIMA model has been used as the technique. Our final objective is to analyze retweets comments and tweets on particular two products. Output is displayed as a graph. Propagation topology is used as the technique for retweet analysis and exponential regression function is used for popularity prediction.

Keywords

Twitter; Sentimental Analysis; Machine Learning; Clustering; Graph Mining; Data Mining.

1. Sentiment analysis https://en.wikipedia.org/wiki/Sentiment_analysis [Accessed: Jul. 19, 2017.]


2. Predicting the Future with Social Media http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=5616710&tag=1 .


3. Analyzing and predicting news popularity on Twitter, B Wu, H Shen. Department of Electrical and Computer Engineering, Clemson University, United States.


4. Logistic Regression for Machine Learning http://machinelearningmastery.com/logistic-regression-for-machine-learning/


5. How Naive Bayes classifier algorithm works in machine learning http://courseweb.sliit.lk/pluginfile.php/76746/mod_resource/content/1/eguid.pdf


6. Natural Language Toolkit . http://www.nltk.org/. [Accessed: Jul. 26, 2017.]


7. Stop words with NLTK https://pythonprogramming.net/stop-wordsnltk-tutorial/ [8]time analysis and forecasting algorithms [Online] Available https://www.slideshare.net/tharindurusira/time-series-prediction-algorithms-literature-review


8. TheARIMAProcedure, http://www.dms.umontreal.ca/~duchesne/chap7.pdf


9. Univariate Time Series Analysis; ARIMA Models http://www.econ.ku.dk/metrics/Econometrics2_05_II/Slides/07_univariatetimeseries_2pp.pdf


10. Supervised k-Means Clustering http://www.cs.cornell.edu/~tomf/publications/supervised_kmeans-08.pdf


11. python-cluster Documentation. https://media.readthedocs.org/pdf/python-luster/latest/python-cluster.pdf