Thumbs Up or Thumbs Down? Semantic Orientation Applied to Unsupervised Classification of Reviews

Download

Skip this Video

Loading SlideShow in 5 Seconds..

Thumbs Upwardly or Thumbs Down? Semantic Orientation Practical to Unsupervised Nomenclature on Reviews PowerPoint Presentation

Download Presentation

Thumbs Up or Thumbs Down? Semantic Orientation Applied to Unsupervised Classification on Reviews

Thumbs Up or Thumbs Down? Semantic Orientation Applied to Unsupervised Classification on Reviews

Download Presentation

Thumbs Up or Thumbs Downwards? Semantic Orientation Applied to Unsupervised Nomenclature on Reviews

- - - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - - -

Presentation Transcript

---

1. Thumbs Up or Thumbs Down? Semantic Orientation Applied to Unsupervised Classification on Reviews Peter D. Turney Found for Information Technology National Research Council of Canada Ottawa, Ontario, Canada, K1A 0R6 peter.turney@nrc.ca Proceedings of the 40th Annual Coming together of the Association for Computational Linguistics (ACL), Philadelphia, July 2002, pp. 417-424.

2. ane. Introduction • If yous are considering a holiday in Akumal, United mexican states, y'all might go to a search engine and enter the query "Akumal travel review". Google reports about five,000 matches. Information technology would exist useful to know what fraction of these matches recommend Akumal as a travel destination • Other application: Recognizing flames, Developing new kinds of search tools

3. 1. Introduction • This paper present a elementary unsupervised learning algorithm for classifying a review as recommended or non recommended • Input: written review, Output: nomenclature • Using POS tagger to identify phrases in the input text that incorporate adjectives or adverbs • Estimating the semantic orientation of each extracted phrase • Assigning the given review to a class, recommended or not recommended, based on the average semantic orientation of the phrases extracted from the review

4. ane. Introduction • The PMI-IR algorithm is employed to estimate the semantic orientation of a phrase (Turney, 2001) • PMI-IR uses Pointwise Common Information (PMI) and Information Retrieval (IR) to measure the similarity of pairs of words or phrases. • The semantic orientation of a given phrase is calculated by comparing its similarity to a positive reference word "first-class" with its similarity to a negative reference word "poor"

5. 2. Classifying Reviews • Past work has demonstrated that adjectives are proficient indicators of subjective, evaluative sentences (Hatzivassiloglou & Wiebe, 2000; Wiebe, 2000; Wiebe et al., 2001). • However, although an isolated describing word may betoken subjectivity, there may exist insufficient context to decide semantic orientation. • Ex: unpredictable, simple etc.

6. 2. Classifying Reviews • The offset stride a POS tagger is applied to the review (Brill, 1994) Two consecutive words are extracted from the review if their tags conform to any of the patterns

7. 2. Classifying Reviews • The second step is to estimate the semantic orientation of the extracted phrases, using the PMI-IR algorithm. This algorithm uses mutual information as a measure of the strength of semantic clan betwixt two words (Church & Hanks, 1989).

8. 2. Classifying Reviews • The Pointwise Common Information (PMI) between ii words, word1 and word2, is divers every bit follows (Church & Hanks, 1989): • SO(phrase) = PMI(phrase, "splendid") – PMI(phrase, "poor") (2)

9. 2. Classifying Reviews • PMI-IR estimates PMI by issuing queries to a search engine and noting the number of hits (matching documents). The following experiments employ the AltaVista Advanced Search engine, which indexes approximately 350 million web pages • Choosing AltaVista because information technology has a Almost operator. The AltaVista Nearly operator constrains the search to documents that contain the words within ten words of one another, in either order.

10. 2. Classifying Reviews • To avoid division by zero, adding 0.01 to the hits • Too skipped phrase when both hits(phrase Most "fantabulous") and hits(phrase Virtually "poor") were less than four

11. 2. Classifying Reviews • The third step is to calculate the average semantic orientation of the phrases in the given review and classify the review as recommended if the average is positive and otherwise not recommended

12. ii. Classifying Reviews

13. two. Classifying Reviews

14. 3. Related Work • This work is most closely related to Hatzivassiloglou and McKeown's (1997) piece of work on predicting the semantic orientation of adjectives. They note that in that location are linguistic constraints on the semantic orientations of adjectives in conjunctions

15. 3. Related Piece of work • The tax proposal was simple and well received by the public. • The tax proposal was simplistic just well received by the public. • (*) The taxation proposal was simplistic and well received by the public.

16. three. Related Work • All conjunctions of adjectives are extracted from the given corpus. • A supervised learning algorithm combines multiple sources of evidence to label pairs of adjectives as having the same semantic orientation or different semantic orientations. The result is a graph where the nodes are adjectives and links indicate sameness or difference of semantic orientation.

17. three. Related Work • A clustering algorithm processes the graph structure to produce two subsets of adjectives, such that links across the two subsets are mainly different-orientation links, and links inside a subset are mainly aforementioned-orientation links • Since it is known that positive adjectives tend to exist used more frequently than negative adjectives, the cluster with the higher average frequency is classified as having positive semantic orientation. • This algorithm classifies adjectives with accuracies ranging from 78% to 92%, depending on the amount of grooming data that is available.

18. iii. Related Work • Other related work is concerned with determining subjectivity (Hatzivassiloglou & Wiebe, 2000; Wiebe, 2000; Wiebe et al., 2001). • The task is to distinguish sentences that present opinions and evaluations from sentences that objectively nowadays factual information (Wiebe, 2000) • A variety of potential applications for automatic subjectivity tagging, such as recognizing "flames" (Spertus, 1997), classifying e-mail, recognizing speaker role in radio broadcasts, and mining reviews.

19. four. Experiments

20. 4. Experiments

21. 5. Discussion of Results

22. 5. Discussion of Results

23. five. Discussion of Results • A limitation of PMI-IR is the fourth dimension required to send queries to AltaVista. Inspection of Equation (3) shows that information technology takes four queries to calculate the semantic orientation of a phrase

24. 6. Applications • Providing summary statistics for search engines. Given the query "Akumal travel review", a search engine could report, "There are 5,000 hits, of which 80% are thumbs upwards and 20% are thumbs downward." • Filtering "flames" for newsgroups (Spertus, 1997).

25. 7. Conclusions • Movie reviews are difficult to classify, considering the whole is non necessarily the sum of the parts • On the other mitt, for banks and automobiles, information technology seems that the whole is the sum of the parts • The simplicity of PMI-IR may encourage farther work with semantic orientation. • The limitations of this work include the fourth dimension required for queries and, for some applications, the level of accuracy that was achieved.

feywornibrink.blogspot.com

Source: https://www.slideserve.com/clancy/thumbs-up-or-thumbs-down-semantic-orientation-applied-to-unsupervised-classification-on-reviews

Share this post