We describe a sentiment classification method that is applicable when we do not have any labeled data for a target domain but have some labeled data for multiple other domains, designated as the source domains.

Unlike previous cross-domain sentiment classification methods, our method can efficiently learn from multiple source domains.

Our method significantly outperforms numerous baselines and returns results that are better than or comparable to previous cross-domain sentiment classification methods on a benchmark dataset containing Amazon user reviews for different types of products.

Automatic document level sentiment classification (Pang et al., 2002; Tumey, 2002) is the task of classifying a given review with respect to the sentiment expressed by the author of the review.

For example, a sentiment classifier might classify a user review about a movie as positive or negative depending on the sentiment

Sentiment classification has been applied in numerous tasks such as opinion mining (Pang and Lee, 2008), opinion summarization (Lu et al., 2009), contextual advertising (Fan and Chang, 2010), and market analysis (Hu and Liu, 2004).

Furthermore, using feature augmentation and selection according to the information gain criteria for cross-domain sentiment classification , our proposed approach performs either better or comparably compared to previous approaches.

Given a piece of text, sentiment classification aims to determine whether the semantic orientation of the text is positive, negative or neutral.

With prior polarity words extracted from both the MPQA subjectivity lexicon1 and the appraisal lexiconz, the J ST model achieves a sentiment classification accuracy of 74% on the movie review data3 and 71% on the multi-domain sentiment dataset4.

The fact that the J ST model does not required any labeled documents for training makes it desirable for domain adaptation in sentiment classification .

These observations motivate us to explore polarity-bearing topics extracted by JST for cross-domain sentiment classification since grouping words from different domains but bearing similar sentiment has the effect of overcoming the data distribution difference of two domains.

Output: A sentiment classifier for the target domain Pt 1: Merge D8 and Pt with document labels discarded, D: {(51331 3 ng Nfimfwl 3 ng Nt}

Input: The target domain data, 13’5 = E X : 1 g n g N75, N75 > N5}, document sentiment classification threshold 7' Output: A labeled document pool 13 1: Train a J ST model parameterized by A on D75 2: for each document 533,5, 6 D75 do 3: Infer its sentiment class label from JST as ln = arg max, P(l|:cf,; A)

proposed structural correspondence learning (SCL) for domain adaptation in sentiment classification .

In this paper, we focus on target-dependent Twitter sentiment classification ; namely, given a query, we classify the sentiments of the tweets as positive, negative or neutral according to whether they conuun posfljve, negafive or nqual senfi-ments about that query.

However, because tweets are usually short and more ambiguous, sometimes it is not enough to consider only the current tweet for sentiment classification .

In this paper, we propose to improve target-dependent Twitter sentiment classification by 1) incorporating target-dependent features; and 2) taking related tweets into consideration.

3r Sentiment Classification

The problem needing to be addressed can be formally named as Target-dependent Sentiment Classification of Tweets; namely, given a query, classifying the sentiments of the tweets as positive, negative or neutral according to whether they contain positive, negative or neutral sentiments about that query.

The state-of-the-art approaches for solving this problem, such as (Go et al., 20095; Barbosa and Feng, 2010), basically follow (Pang et al., 2002), who utilize machine learning based classifiers for the sentiment classification of texts.

Given the input data D1, D2 and U, our task is to jointly learn two monolingual sentiment classifiers — one for L1 and one for L2.

We present a novel approach for joint bilingual sentiment classification at the sentence level that augments available labeled data in each language with unlabeled parallel data.

We rely on the intuition that the sentiment labels for parallel sentences should be similar and present a model that jointly learns improved monolingual sentiment classifiers for each language.

Not surprisingly, most methods for sentiment classification are supervised learning techniques, which require training data annotated with the appropriate sentiment labels (e. g. document-level or sentence-level positive vs. negative polarity).

In addition, there is still much room for improvement in existing monolingual (including English) sentiment classifiers , especially at the sentence level (Pang and Lee, 2008).

In contrast to previous work, we (1) assume that some amount of sentiment-labeled data is available for the language pair under study, and (2) investigate methods to simultaneously improve sentiment classification for both languages.

Prettenhofer and Stein (2010) investigate cross-lingual sentiment classification from the perspective of domain adaptation based on structural correspondence learning (Blitzer et al., 2006).

Approaches that do not explicitly involve resource adaptation include Wan (2009), which uses co-training (Blum and Mitchell, 1998) with English vs. Chinese features comprising the two independent “views” to exploit unlabeled Chinese data and a labeled English corpus and thereby improves Chinese sentiment classification .

Unlike the methods described above, we focus on simultaneously improving the performance of sentiment classification in a pair of languages by developing a model that relies on sentiment-labeled data in each language as well as unlabeled parallel text for the language pair.

We show that this constraint is effective on the sentiment classification task (Fang et al., 2002), resulting in scores similar to the ones obtained by the structural correspondence methods (Blitzer et al., 2007) without the need to engineer auxiliary tasks.

Our approach results in competitive domain-adaptation performance on the sentiment classification task, rivalling that of the state-of-the-art SCL method (Blitzer et al., 2007).

In this section we empirically evaluate our approach on the sentiment classification task.

On the sentiment classification task in order to construct them two steps need to be performed: (1) a set of words correlated with the sentiment label is selected, and, then (2) prediction of each such word is regarded a distinct auxiliary problem.

We evaluate our approach on adapting sentiment classifiers on 4 domains: books, DVDs, electronics and kitchen appliances (Blitzer et al., 2007).

Research has been made on sentiment classification in document-level (Turney et al., 2002, Pang et al., 2002, Seki et al.

The discourse of contentious issues in news articles show different characteristics from that studied in the sentiment classification tasks.

They assume a debate frame, which is similar to the frame of the sentiment classification task, i.e., for vs. against the debate topic.

Research on sentiment classification and debate stance recognition takes a topic-oriented view, and attempts to perform classification under the ‘positive vs. negative’ or ‘for vs. against’ frame for the given topic, e. g., positive vs. negative about iPhone.

We evaluate the model using small, widely used sentiment and subjectivity corpora and find it outperforms several previously introduced methods for sentiment classification .

Of course, this is only an impressionistic analysis of a few cases, but it is helpful in understanding why the sentiment-enriched model proves superior at the sentiment classification results we report next.

Martineau and Finin present evidence that this weighting helps with sentiment classification , and Paltoglou and Thelwall (2010) systematically explore a number of weighting schemes in the context of sentiment analysis.