Classifying with Co-stems

Lipka, Nedim; Stein, Benno

doi:10.1007/978-3-642-20161-5_31

Cited by 3 publications

(3 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…k-top co-stems. In prior work, the ends of words (e.g., conjugations, plurals, and possessive marks) were shown to give notable signal about a variety of blog author attributes (Lipka and Stein 2011). These strings, called co-stems, can be obtained by subtracting the stem returned by the Lovins stemmer and processing only the ending that remains (i.e., the word minus the stem).…”

Section: Featuresmentioning

confidence: 99%

Homophily and Latent Attribute Inference: Inferring Latent Attributes of Twitter Users from Neighbors

Zamal

Liu

Ruths

2021

ICWSM

View full text Add to dashboard Cite

In this paper, we extend existing work on latent attribute inference by leveraging the principle of homophily: we evaluate the inference accuracy gained by augmenting the user features with features derived from the Twitter profiles and postings of her friends. We consider three attributes which have varying degrees of assortativity: gender, age, and political affiliation. Our approach yields a significant and robust increase in accuracy for both age and political affiliation, indicating that our approach boosts performance for attributes with moderate to high assortativity. Furthermore, different neighborhood subsets yielded optimal performance for different attributes, suggesting that different subsamples of the user's neighborhood characterize different aspects of the user herself. Finally, inferences using only the features of a user's neighbors outperformed those based on the user's features alone. This suggests that the neighborhood context alone carries substantial information about the user.

show abstract

Section: Featuresmentioning

confidence: 99%

Homophily and Latent Attribute Inference: Inferring Latent Attributes of Twitter Users from Neighbors

Zamal

Liu

Ruths

2021

ICWSM

View full text Add to dashboard Cite

show abstract

“…A similar procedure was used to select the k mostbiased stemmed tokens and hashtags for each class. Finally, we include the k most-biased co-stems, which proved useful for inferring author attributes (Lipka and Stein 2011). The same value of k=13 was used for all four feature types after initial tuning showed it provided good performance; higher values of k did not noticeably improve performance.…”

Section: Methodsmentioning

confidence: 99%

Organizations Are Users Too: Characterizing and Detecting the Presence of Organizations on Twitter

McCorriston

Jurgens

Ruths

2021

ICWSM

View full text Add to dashboard Cite

Much work on the demographics of social media platforms such as Twitter has focused on the properties of individuals, such as gender or age. However, because credible detectors for organization accounts do not exist, these and future large-scale studies of human behavior on social media can be contaminated by the presence of accounts belonging to organizations. We analyze organizations on Twitter to assess their distinct behavioral characteristics and determine what types of organizations are active. We first create a dataset of manually classified accounts from a representative sample of Twitter and then introduce a classifier to distinguish between organizational and personal accounts. In addition, we find that although organizations make up less than 10% of the accounts, they are significantly more connected, with an order of magnitude more friends and followers.

show abstract

“…They consist feature male female words team, league, bro hair, omg, cute stems scor, win, team girl, bab, feel co-stems s, e, ed n, i, a digrams in, er, re ee, ah, aa trigrams ing, ion, ent aha, eee, aaa hashtags soundcloud, mufc, np sorrynotsorry, excited, love of the k-top most discriminating words, k-top word stems 9 (e.g., "paper" representing "paper", "papers", or "papered"). k-top co-stems (Lipka and Stein [13] demonstrate that the stem-reduced words, or co-stems, yield a significant improvement over classical bag of words models; for example, "s" for "papers"), k-top digrams (the most discriminating digrams; for example, "pa", "ap", "pe", and "er"), k-top trigrams (the most discriminating trigrams; for example, "pap", "ape", and "per"), and k-top hashtags (hashtags are labels that are attached by the users to their Tweet messages; for example, "paper" for "#paper").…”

Section: Threshold Classifiermentioning

confidence: 99%

Gender Inference using Statistical Name Characteristics in Twitter

Mueller

Stumme

2016

Proceedings of the the 3rd Multidisciplinary International Social Networks Conference on SocialInformatics 2016, Data Science 2

View full text Add to dashboard Cite

Much attention has been given to the task of gender inference of Twitter users. Although names are strong gender indicators, the names of Twitter users are rarely used as a feature; probably due to the high number of ill-formed names, which cannot be found in any name dictionary. Instead of relying solely on a name database, we propose a novel name classifier. Our approach extracts characteristics from the user names and uses those in order to assign the names to a gender. This enables us to classify international first names as well as ill-formed names.

show abstract

Classifying with Co-stems

Cited by 3 publications

References 17 publications

Homophily and Latent Attribute Inference: Inferring Latent Attributes of Twitter Users from Neighbors

Homophily and Latent Attribute Inference: Inferring Latent Attributes of Twitter Users from Neighbors

Organizations Are Users Too: Characterizing and Detecting the Presence of Organizations on Twitter

Gender Inference using Statistical Name Characteristics in Twitter

Contact Info

Product

Resources

About