1. Introduction

In the globalized information age, social media has become one of the main channels for people to obtain news and information. As a result of social media tools, geographic barriers that once restricted communication have been broken down and e-participation, virtual presence, and a diversity of online communities have been transformed [1]. Especially in times of geopolitical-military conflict, social media has come to the fore as a platform for the rapid dissemination of news, views, and analysis.

Geopolitical conflicts usually have a significant impact on the local and international community due to the urgency and international reach of the conflict, which stirs up public emotions. As the term implies, conventional conflict is a contested incompatibility between two parties in terms of government and/or territory, in which armed forces are used between them [2]. There are three types of wars: (1) minor armed conflicts (0-25 deaths), (2) intermediate armed conflicts (25-1000 deaths), and (3) wars (1000+ deaths) [2]. There are four types of actors: (1) extra systemic, (2) interstate, (3) internal, and (4) international [2]. This research focuses on Israeli-Palestinian which is an international war.

Commencing in 1948, the Israeli-Palestinian conflict is intricately linked to the Middle East's wars as well as the Arab-Israeli conflict involving territorial, ethnic, religious, and political disputes. Thus, the Israeli-Palestinian conflict is a big, intricate, and challenging issue [3]. On October 7, 2023, the Palestinian Islamic Resistance Movement launched massive rocket attacks on Israel, and a new round of the Palestinian-Israeli conflict began. Until April, 2024, a new phase of the Israeli-Palestinian war has lasted for half a year. The ceasefire talks have come to a standstill, the humanitarian crisis has gotten worse, and the armed combat between Israel and the Palestinians in the Gaza Strip has not stopped. Furthermore, the situation in the Israeli-Palestinian territory has been worse since 2023. Clashes between the Israeli army and the Palestinians have persisted, and the Israeli army has conducted multiple search and arrest operations in the West Bank. In addition, there have been several clashes between the two sides in Jerusalem, the Gaza Strip, and other locations, which have left many people dead. Discussions on the Israel-Palestine issue are among the topics frequently addressed by users on the Chinese social media platform Weibo.

Weibo is a platform for sharing, disseminating, and accessing information based on user relationships, founded on 14 August 2009. Then it is one of the largest social media platforms in China, apart from Twitter-like functionalities, microblogging offers threaded comments, applications, games, and microblogging medals and it also permits the uploading of rich media in user feeds [4]. Its daily active users reached 46.2 million, with users posting more than 100 million blogs per day. Because of its rich functions, thousands of Chinese users of different ages, genders, and occupations can post their opinions on different events on it every day. In the past, Chinese people on Weibo have provided different and varied attitudes and sentiments in the evaluation of geopolitical conflicts. As a result, Weibo has become a major channel for Chinese users to express their views on the Israeli-Palestinian conflict.

2. Literature Review

As social media has become more and more prominent in our lives, a number of scholars have begun to analyse it. In Alalwan‘s research, they found that the number of studies in the field of social media marketing has increased dramatically in recent years [5]. And in analysing comments of social media, Lin et al use Natural Language Processing (NLP) techniques to thereby mine, identify and reveal the multi-dimensional risk of consumers' perceptions of online gaming on social media [6]. Social media comment analytics has applications not only in the behaviour of users, but also in the different content of their comments. By analysing the Jordanian population's use of humorous comments to cope with and ridicule the outbreak during the COVID-19 pandemic, Alkaraki and Alias found that humorous comments covered a variety of topics such as the government, gender, COVID-19, embargoes, behaviours, conspiracy theories, geographic areas, and masks [7].

While not much research has been done on Weibo, there is a lot of research on other specific social media. After analysing information on Facebook about vaccine errors, Klimiuk et al got information about a possible lack of trust in the achievements of medical science on the part of the people posting the comments [8]. In their research, one limitation is that comment content is hard to clarify and the boundaries between categories were not always clear. And another significant limitation of this study is that its analysis was limited to a single social media and did not extend to a broader range of content on Facebook or other online comments. Through multi-platform comments analysis, the overall data can be better obtained and the conclusions obtained will be more accurate.

In the social media, there are some research about politics. Kriaučiūnienė et al analysed comments on Facebook and Twitter about US presidential candidates. The authors used the topic of the 2016 Presidential Election in the United States to interpret linguistic as well as other multimodal means of expressing stances that were used by users of social networks in their writing spaces. In terms of the emotional bent of the content of the comments, they observed that the comments about Trump and Clinton were roughly balanced on the positive and negative sides. While comments below Donald Trump's posts tended to be supportive and Hillary Clinton's posts received more positive feedback, a deeper analysis of user comments revealed that Donald Trump used more negative words and phrases. By contrast, Hillary Clinton mostly tries to stay neutral or positive in her posts [9].

Another focus on the theme is Geopolitical conflict. In Hansan et al’s research, their topic is about Russia–Ukraine War. They present a study about sentiment analysis in Bengali, specifically Bengali commentary on the ongoing conflict between Russia and Ukraine. Due to a shortage of datasets in Bengali, which is one of the most widely spoken languages in the world, they created a dataset made up of Bangladeshi comments on the ongoing Russian-Ukrainian war, thus demonstrating the advantages of the Transformer model for natural language processing tasks [10].

While the research and ideas mentioned above aid in understanding of certain underlying knowledge beforehand. There are still existing research gaps. Although there are many studies on geopolitical conflicts and social media, few articles have been written on the discursive analyses of the Palestinian-Israeli conflict on Weibo. This study is very important to analyze and understand Chinese people's attitudes towards the Palestinian-Israeli conflict. The study on the behavior of micro-blog users in the Palestinian-Israeli conflict is helpful to further understand the Chinese public's attitude and emotional response to the conflict, and further understand China's position and public opinion orientation on international geopolitical issues, as well as the important significance of social media public opinion monitoring, international relations research, and public opinion guidance.

3. Methodology

This section proposes the idea of using content analysis to conduct sentiment statistics and word frequency analysis on comments on Weibo. In the collection of data for the analysis of emotional content, search for specific topics or keywords, and select content with 5,000 comments that are highly liked and forwarded on Weibo. For a particular topic, select a topic with more than 3,000 likes and more than 100 comments. And delete comments that have nothing to do with the topic. Then the researcher used ERNIE Bot, which is a technology based on natural language processing and emotion analysis, to identify the emotions in the text. The researcher input the selected 5000 comments into ERNIE Bot for emotion analysis, and identify the positive, negative, and neutral emotions among them. In order to ensure the accuracy of the results of emotion analysis, the researcher carried out manual verification to check and correct the emotions identified by the ERNIE Bot.

When analyzing the word frequency of comments, according to the long tail law, the researcher chose 5,000 Weibo comments with 0 likes and retweets for word frequency analysis. These comments often reflect diverse perspectives and emotional expressions on social media, contributing to a comprehensive understanding of discussions about the Israeli-Palestinian conflict. Search and filter the comments with 0 likes and forwards in different time periods through Weibo to ensure that the selected comments have a certain. Then the researcher input the selected 5,000 comments into ERNIE Bot for word frequency analysis. ERNIE Bot can identify and count the words that appear more frequently in the text. In order to avoid the interference of common words, I carried out stop word filtering before word frequency analysis to filter out some common meaningless words and only retain the words with actual meaning. And get the top 4 most used words.

4. Results

Use Microsoft Excel to create a line chart to visualize the trend of different emotions over time. The horizontal axis of the line chart represents time, the vertical axis represents the proportion of different emotions, and different colored lines represent different emotions.

Figure 1: Sentiment Analysis of Social Media Users on Geopolitical Conflict(October 2023 – April 2024)

According to Figure 1, the proportion of negative feelings increased significantly from 33.33% in October 2023 to 100% in April 2024. This trend shows that public discontent and concern about the conflict has grown steadily over time, reaching a peak in early 2024. Positive feelings decreased from 47.62% in October 2023 to 0% in April 2024. This sharp decline indicates that public hopes for a peaceful resolution of the conflict have diminished significantly as the conflict has dragged on. Neutral sentiment experienced fluctuations, initially at 19.05% in October 2023, then dropping to 0% by April 2024.

Use Microsoft Excel to create a heatmap to visualize the proportion of different words used over time. The vertical axis of the heatmap represents time, the horizontal axis represents words, and the color intensity indicates the proportion of words used, with darker colors indicating higher proportions.

Table 1: Frequency of Top Words in Geopolitical Conflict by Month

In Table 1, the high frequency of "world peace" and "hope" from October 2023 (34.78% each) reflects the high public expectations for peace at the beginning of the conflict. The increased frequency of "ceasefire" and "hope" in November 2023 (23.53% and 17.65%, respectively) . In December 2023, the word "peace" dominated (47.06%). However, starting in January 2024, references to "Israel" and "Hamas" increased. In February, "Israel" and "Hamas" each had 36.36 percent. In March 2024, references to "Israel" and "children" increased. In April 2024, the words "Israel" and "Hamas" were mentioned frequently, and the words "disaster" appeared.

5. Discussion

According to the graph obtained in results, it shows that he increase in negative emotions indicates growing public discontent and concern about the ongoing conflict, peaking in early 2024. This may be because the conflict is likely to intensify in late 2023 and early 2024, leading to more casualties and destruction and triggering public outcry and concern. The escalation of violence and conflict is likely to be the main reason for the increase in negative sentiment. The second reason is that the continuation of the conflict could lead to serious humanitarian crises, including refugee problems, and shortages of food and medical resources. These problems may exacerbate public concern and dissatisfaction with the conflict. Humanitarian crises often lead to widespread social discontent because of their direct threat to people's basic survival needs and dignity. The speed of response of the government and relevant agencies, the fairness of resource allocation and the transparency of information are all important factors affecting public satisfaction. In responding to such crises, these factors need to be considered in order to reduce social discontent and maintain social stability. The upsurge in negative sentiment could lead to increased discussion of the Israeli-Palestinian conflict on social media, with more members of the public and opinion leaders participating in the discussion. Besides, public attention to the Israeli-Palestinian conflict may increase overall interest and engagement in international affairs, driving more people to pay attention to and discuss global issues.

At the same time, positive and neutral statements began to decline, possibly because as the conflict lasted longer and hopes for a peaceful resolution of the conflict faded, the public became disillusioned and frustrated, leading to a decrease in positive statements, and the intensification of the conflict and the increase in violence, such as civilian casualties and the destruction of infrastructure, would further reduce the positive sentiment of the public. Discussions on social media are often emotional, and negative emotions are more likely to spread and amplify, leading to a decrease in positive and neutral comment. With fewer positive comments on social media, negative emotions are easy to spread quickly on social media, and lead to more negative emotions, forming a kind of "emotional contagion effect", resulting in general low public mood. In Kramer et al’s research, it shows that experiments on reducing emotional content in news feeds on Facebook have found that when positive expressions are reduced, people have fewer positive posts and more negative posts, and vice versa [11]. Moreover, it is proposed that social media platforms can encourage the spread of more positive content and reduce the spread of negative emotions through algorithm adjustment and content review [11].

On the Weibo, the key word in people's initial comments on the Israeli-Palestinian conflict was generally peace. On social media, there can be a lot of coverage of the United Nations and other international organizations, national mediation efforts with the parties to the conflict, and ceasefire initiatives. These messages encourage public discussion of peace and the hope of ending conflicts through peaceful means. In the early stages of a conflict, there is often a certain optimism among the public that diplomacy and negotiation can quickly resolve the conflict and bring about peace. Besides, in terms of social media, in the early stages of a conflict, media coverage of international calls and efforts for a peaceful resolution of the conflict often influences the direction of public discussion. At the same time, social media platforms may prioritize content that calls for peace, creating a mainstream atmosphere for peaceful discussion which can promote rational dialogue and constructive discussion and reduce the spread of violence and antagonistic comment.

Initially, there was a focus on peace and ceasefire, but as the conflict continued, the discourse shifted towards specific entities involved ("Israel" and "Hamas") and the humanitarian impact. The reasons for this shift may be multifaceted. First, when a conflict begins, people may first focus on how to achieve peace and a ceasefire to quickly end the violence. However, as the conflict continues and potentially escalates, public and media attention may turn to specific entities involved in the conflict, such as "Israel" and "Hamas," as the actions and tactics of these entities directly affect the development and outcome of the conflict. The second aspect may be because the media tend to adjust the Angle of reporting conflict according to the progress of the event and the attention of the audience. Early in the conflict, coverage may focus more on overall peace calls and ceasefire efforts. But as the conflict deepens, the media may focus more on specific parties to the conflict and affected populations, leading to a shift in public attention. Overall, this shift in discourse from peace and ceasefire to concrete entities and humanitarian impact has increased public awareness of the conflict and increased attention to humanitarian crises.

6. Conclusion

This analysis reveals trends in public sentiment towards the Israeli-Palestinian conflict. In April 2024, there is a significant increase in negative sentiment to 100 percent, indicating growing public dissatisfaction and concern. This increase is closely linked to the escalation of the conflict, increased casualties and the humanitarian crisis, which have fuelled public discontent. Negative sentiment is further amplified on social media, creating a "contagion effect" that leads to overall low public sentiment. In contrast, positive sentiment drops sharply to 0 percent by April 2024, reflecting fading hopes for a peaceful resolution of the conflict. This decline may be attributed to the protracted nature of the conflict and increased violence, eroding public optimism. Neutral sentiment has also declined, suggesting that public discussion has become more polarised and emotional. Changes in the use of keywords also reflect a shift in public and media focus. There is a shift from "world peace" and "hope" to specific entities such as "Israel" and "Hamas" and "disaster". as well as terms such as "catastrophe", suggests a shift in public discourse from peace and ceasefire efforts to specific participants in the conflict and their humanitarian impact. This shift may be due to the continuation of the conflict and a shift in media coverage from overall calls for peace to specific parties to the conflict and affected populations. These findings emphasise the importance of effectively addressing public grievances and managing humanitarian crises. Social media platforms can moderate the spread of negative sentiment by promoting positive content and creating inclusive, impartial narratives. Understanding these dynamics can help create a more equitable and engaging environment for public discourse, which ultimately contributes to social stability and peace efforts. While the study analyzed comments about the Israeli-Palestinian conflict on the Chinese social media site Weibo, there are some limitations. Comments on Weibo do not represent the views of all Chinese citizens, as there are other social media platforms in China, such as WeChat and QQ. Therefore, based on this, future research could be conducted on other Chinese social media platforms to analyze news and commentary related to geopolitical conflicts to obtain a more comprehensive public view.