Examining Crisis Communication Using Semantic Network and Sentiment Analysis: A Case Study on NetEase Games

We generated a “one-dimensional co-occurrence matrix” using the 50 most important keywords and estimated modularity to gauge the quality of the communities by calculating the term frequency-inverse document frequency (TF-IDF). The TF-IDF method is a statistical analytic method for determining the relative relevance of keywords in document sets or corpora. For example, if a word is somewhat uncommon yet appears frequently in a certain article, it will most likely represent the qualities of the piece.

The topic model obtained by LDA has been used for representing text collections as in [58, 122, 123]. The second most used source is Wikipedia [73], which covers a wide range of subjects and has the advantage of presenting the same concept in different languages. Wikipedia concepts, as well as their links and categories, are also useful for enriching text representation [74–77] or classifying documents [78–80]. Bos [31] presents an extensive survey of computational semantics, a research area focused on computationally understanding human language in written or spoken form.

The Power of Data Science in Determining Insurance Premiums

Such an algorithm relies exclusively on machine learning techniques and learns on received data. In some cases, this makes customer service far more attentive and responsive, as the customer support team is informed in real-time about any negative comments. Fourth, the relationship between social media content (Taylor et al., 2020), users’ emotions and demands, and mobile game addiction in terms of semantic network and sentiment analysis has not been covered (Mircică, 2020; Pop et al., 2021). Consequently, in order to improve text mining results, many text mining researches claim that their solutions treat or consider text semantics in some way. However, text mining is a wide research field and there is a lack of secondary studies that summarize and integrate the different approaches. Looking for the answer to this question, we conducted this systematic mapping based on 1693 studies, accepted among the 3984 studies identified in five digital libraries.

• Machine learning-based semantic analysis involves sub-tasks such as relationship extraction and word sense disambiguation.
• When you read news about natural disasters, you usually feel negative emotions, and when you read news about the last championship won by your favorite football team, you usually feel positive emotions.
• Interested in building tools that intelligently tracking how interviewees feel about certain topics?
• For more information about NLP and other data analytics processes, reach out to us at
• As, the proposed SALOM model uses the semantic similarity to avoid spam product aspects.
• However, the nature of online comments affects the performance of the opinion mining process because they may contain negation words or unrelated aspects to the product.

The predicted result is visualized based on the color shades, where the darker the color indicates the more samples classified into that category. As illustrated in Figure 5, the color of the squares on the diagonal of our method is darker than that of HLCs method, which means our method outperforms the HLCs and shows more ability to distinguish between various emotion categories. In the FI dataset, the HLCs and our method have less incorrectness on amusement, sadness, contentment, and awe emotions, while the HLCs tends to incorrectly classify the anger and fear emotions.

Sentiment Analysis vs. Semantic Analysis: What Creates More Value?

“New servers” and “attack servers” refer to groups of players threatening or planning to attack the game’s servers, and “closed comments” refer to IC removing negative user comments. Term frequency–inverse document frequency scores of responses to the apology announcements. After several comparisons, the precise mode was selected for this study because the omnibus mode does not resolve ambiguity and the search engine mode is more suitable for search engine analysis. Search engine mode splits longer sentences into smaller ones to improve the recall rate for search engine segmentation. Hence, we replaced those and user-added emoticons and symbols with emotion-colored textual statements to neutralize the non-textual data as much as possible. For the classification part, I developed an algorithm following a Topic Modeling approach as well as a heuristics-based approach that relies on language structure.

It is also a key component of several machine learning tools available today, such as search engines, chatbots, and text analysis software. Sentiment analysis has a good application effect in crisis communication, as it directly presents the two sides of social media users’ comments in the form of scores. For example, Bygstad and Presthus (2013) used Facebook to study customer communication cases of two Scandinavian airlines during the volcanic-ash crisis of April 2010.

Commercial Products:

It is a crucial component of Natural Language Processing (NLP) and the inspiration for applications like chatbots, search engines, and text analysis using machine learning. Figure 3 shows the average recall, precision and f-measure of the three products’ datasets by using the SentiWordnet lexicon and the Subjectivity lexicon for English adjectives. In the above formulas, TP (True-Positive) identifies the number of words that are correctly classified as product aspects. FP (False-Positive) identifies the number of words that are in correctly classified as product aspects.

Most exciting, with semantic analysis you have tremendous latitude about how you approach the analysis. How many times in your life can you say that lack of certainty gives you a leg up? With semantic analysis, you can let the social media speak for itself, revealing to you amazingly accurate and important information that can inform critical decisions.

Basic Units of Semantic System:

The application of natural language processing methods (NLP) is also frequent. Among these methods, we can find named entity recognition (NER) and semantic role labeling. It shows that there is a concern about developing richer text representations to be input for traditional machine learning algorithms, as we can see in the studies of [55, 139–142]. We perform our experiments on five widely used affective datasets for emotion classification and sentiment prediction, including Flickr and Instagram (FI) [9], EmotionROI [35], Flickr [36], Instagram [36], and Twitter [15]. The FI dataset is collected from Flickr and Instagram websites with weakly labelled web images. The EmotionROI dataset is also collected from Flickr, containing 1980 images classified into six emotion categories (anger, disgust, fear, happy, sadness, and surprising).

In addition, the set of semantic concepts constituted by ANPs covers a limited range of semantics. Ahsan et al. [27] proposed to discover event concept attributes from the website and utilize event concepts as the midlevel semantic features to predict the sentiment of event images, but they only focused on images related to events. metadialog.com From the results shown in Table 4, we can observe that our method achieves the best accuracy in all datasets. The prediction accuracy of our method reaches 61.55% on FI dataset and 49.95% on EmotionROI dataset, which surpasses the state-of-art HLCs method by over 6.75% on FI dataset and 7.36% on EmotionROI dataset, respectively.

The Logical Form of Action Sentences

When a user purchases an item on the ecommerce site, they can potentially give post-purchase feedback for their activity. This allows Cdiscount to focus on improving by studying consumer reviews and detecting their satisfaction or dissatisfaction with the company’s products. Semantic

and sentiment analysis should ideally combine to produce the most desired outcome. These methods will help organizations explore the macro and the micro aspects

involving the sentiments, reactions, and aspirations of customers towards a

brand. Thus, by combining these methodologies, a business can gain better

insight into their customers and can take appropriate actions to effectively

connect with their customers.

Players say the Treasure system breaks the game’s balance, widens the gap between money-rich and time-rich players, and makes it harder for ordinary players to enjoy the game. Where [newline]e

[newline](

[newline]C

c [newline]k

[newline],

S [newline])

[newline] is the proportion of two vertices of the community edge in community [newline]

C [newline]c

k [newline]

[newline]. The term after the formula indicates the proportion of at least one vertex of the edge in community [newline]

C [newline]c

k [newline]

[newline]. In a highly modularized network, the connections between nodes are dense, and the connections between nodes in different modules are sparse.

Title: Multilingual Sentiment Analysis Using Latent Semantic Indexing and Machine Learning.

Meanwhile, we compare with a random concept set that randomly selects a given number of concepts from the set of user-generated tags. Additionally, as the value of increases, the accuracy of emotion classification first increases and then decreases. On the one hand, the increase in the size of the concept space will cause difficulties for recognition. On the other hand, the concepts within this space can reasonably infer the semantic information of other concepts by calculating the mutual information. As the best performance is obtained with , we choose it in all our experiments.

Following this thought of the line, we propose a novel affective semantic concepts discovery method by exploiting shared images and corresponding tags for image emotion classification. Unlike other computer vision tasks, visual emotion analysis is subjective and culture-dependent, which suffers from a bigger “affective gap” between low-level visual features and high-level emotional responses. Early researches on this issue extracted the low-level visual features related to emotions (e.g., colors, texture, and shapes) from input images [5–7].

Designing an Email Management Solution Based on Open Source NLU

The FI dataset is split randomly into 80% for training, 5% for validation, and 15% for testing. For the Flickr dataset and Instagram dataset, we randomly sample the same number of images for each class following the same configuration in [42], which are split randomly into around 90% for training and 10% for testing. The remaining datasets are all randomly divided into 80% training set and 20% testing set. Sentiment analysis over Twitter offer organisations a fast and effective way to monitor the publics’ feelings towards their brand, business, directors, etc. A wide range of features and methods for training sentiment classifiers for Twitter datasets have been researched in recent years with varying results.