Sentiment Analysis of Twitter Messages using Word2vec by Weighted Average

Author(s):  
Kamel Ahsene Djaballah ◽  
Kamel Boukhalfa ◽  
Omar Boussaid
2019 ◽  
Vol 1 (1) ◽  
pp. 1-9
Author(s):  
S. Luintel ◽  
R.K. Sah ◽  
B.R. Lamichhane

There is an excessive growth in user generated textual data due to increment in internet and social media users which includes enormous amount of sarcastic words, emoji, sentences. Sarcasm is a nuanced form of communication where individual states opposite of what is implied which is done in order to insult someone, to show irritation, or to be funny. Sarcasm is considered as one of the most difficult problems in sentiment analysis due to its ambiguous nature. Recognizing sarcasm in the texts can promote many sentiment analysis and text summarization applications. So for addressing the problem of sarcasm many steps have been adopted for sarcasm detection. Different preprocessing techniques such as Hypertext markup language removal, stop words removal, etc. have been done. Similarly, conversion of the emoji and smileys into their textual equivalent has been performed. Most frequent features has been selected and a hybrid cascade and hybrid weighted average approaches which are the combinations of the algorithms random forest, naïve Bayes and support vector machine have been used for sarcasm detection. The comparison of these two approaches on different basis has been done which has shown cascade outperformed weighted approach. Moreover, comparison of cascade approaches in terms of the algorithm placement has also been performed in which random forest has proved to be the best.


Author(s):  
Agung Eddy Suryo Saputro ◽  
Khairil Anwar Notodiputro ◽  
Indahwati A

In 2018, Indonesia implemented a Governor's Election which included 17 provinces. For several months before the Election, news and opinions regarding the Governor's Election were often trending topics on Twitter. This study aims to describe the results of sentiment mining and determine the best method for predicting sentiment classes. Sentiment mining is based on Lexicon. While the methods used for sentiment analysis are Naive Bayes and C5.0. The results showed that the percentage of positive sentiment in 17 provinces was greater than the negative and neutral sentiments. In addition, method C5.0 produces a better prediction than Naive Bayes.


Author(s):  
Larisa A. Pautova ◽  
Vladimir A. Silkin ◽  
Marina D. Kravchishina ◽  
Valeriy G. Yakubenko ◽  
Anna L. Chultsova

The structure of the summer planktonic communities of the Northern part of the Barents sea in the first half of August 2017 were studied. In the sea-ice melting area, the average phytoplankton biomass producing upper 50-meter layer of water reached values levels of eutrophic waters (up to 2.1 g/m3). Phytoplankton was presented by diatoms of the genera Thalassiosira and Eucampia. Maximum biomass recorded at depths of 22–52 m, the absolute maximum biomass community (5,0 g/m3) marked on the horizon of 45 m (station 5558), located at the outlet of the deep trench Franz Victoria near the West coast of the archipelago Franz Josef Land. In ice-free waters, phytoplankton abundance was low, and the weighted average biomass (8.0 mg/m3 – 123.1 mg/m3) corresponded to oligotrophic waters and lower mesotrophic waters. In the upper layers of the water population abundance was dominated by small flagellates and picoplankton from, biomass – Arctic dinoflagellates (Gymnodinium spp.) and cold Atlantic complexes (Gyrodinium lachryma, Alexandrium tamarense, Dinophysis norvegica). The proportion of Atlantic species in phytoplankton reached 75%. The representatives of warm-water Atlantic complex (Emiliania huxleyi, Rhizosolenia hebetata f. semispina, Ceratium horridum) were recorded up to 80º N, as indicators of the penetration of warm Atlantic waters into the Arctic basin. The presence of oceanic Atlantic species as warm-water and cold systems in the high Arctic indicates the strengthening of processes of “atlantificacion” in the region.


Corpora ◽  
2019 ◽  
Vol 14 (3) ◽  
pp. 327-349
Author(s):  
Craig Frayne

This study uses the two largest available American English language corpora, Google Books and the Corpus of Historical American English (coha), to investigate relations between ecology and language. The paper introduces ecolinguistics as a promising theme for corpus research. While some previous ecolinguistic research has used corpus approaches, there is a case to be made for quantitative methods that draw on larger datasets. Building on other corpus studies that have made connections between language use and environmental change, this paper investigates whether linguistic references to other species have changed in the past two centuries and, if so, how. The methodology consists of two main parts: an examination of the frequency of common names of species followed by aspect-level sentiment analysis of concordance lines. Results point to both opportunities and challenges associated with applying corpus methods to ecolinguistc research.


2017 ◽  
Vol 4 (3) ◽  
pp. 60-71 ◽  
Author(s):  
Alfredo Fort

Though difficult to ascertain because faith based organizations (FBOs) might keep a low profile, be confused with other non-governmental organizations (NGOs), or survey respondents may not know the nature of facilities attended to, these organizations have a long presence in teaching health personnel and delivering health services in many rural and remote populations in the developing world. It is argued that their large networks, logistics agreements with governments, and mission-driven stance brings them closer to the communities they serve, and their services believed of higher quality than average. Kenya has a long history of established FBOs substantial recent health investment by the government. We aimed to find the quantitative and qualitative contributions of FBOs by analyzing two recent data sources: the live web-based nationwide Master Health Facility List, and the 2010 nationwide Service Provision Assessment (SPA) survey. Using this information, we found that FBOs contribute to 11% of all health facilities’ presence in the country, doubling to 23% of all available beds, indicating their relative strength in owning mid-level hospitals around the country. We also constructed an index of readiness as a weighted average from services offered, good management practices and availability of medicines and commodities for 17 items assessed during the SPA survey. We found that FBOs topped the list of managing authorities, with 70 percent of health facility readiness, followed closely by the government at 69 percent, NGOs at 61 percent and lastly a distant private for profit sector at 50 percent. These results seem to indicate that FBOs continue to contribute to an important proportion of health care coverage in Kenya, and that they do so with a relatively high quality of care among all actors. It would be of interest to replicate the analysis with similar databases for other countries in the developing world.


Sign in / Sign up

Export Citation Format

Share Document