Skip to main content

Turkish SentiNet

Project description

Turkish Sentiment Lexicon (HisNet)

Polarity Lexicons

Exploiting a dictionary-based method necessitates the construction of a specific polarity dictionary in the same language as the data-to-be-analyzed. The reason behind this necessity stems from the improbability of creating a universal polarity dictionary due to both grammatical and cultural asymmetries between languages. For instance, a certain historical event can have positive connotations in one culture and negative connotations in another culture. Thus, it is an essential step to create a language specific polarity dictionary.

The first examples of polarity dictionary work could be found in English. SentiWordNet 1.0, the very first study on English polarity dictionaries, was presented by Esuli and Sebastiani (2006). Considerable research has been conducted to im- prove these resources with the aim of making them more precise. For example, the polarities of the objective words in SentiWordNet have been reassessed by Hung and Lini (2010). SenticNet (Cambria et al., 2014), another well-known dictionary in English, is created by rescoring words based on five different criteria, which are happiness, attention, sensitivity, ability and general polarity. Thus, it is evident that SenticNet is a polarity dictionary that provides a more extensive emotional evaluation than SentiWordNet.

There are polar dictionaries created in major languages other than English. However, these dictionaries were found to be insufficient in terms of the number of words. Brooke et al. (2009) aimed to translate English polarity sources to Spanish. At first, the methods established independent from the target language were found adequate, yet in the long term it was noticed that these methods were costly and inaccurate. Employing language-dependent resources to improve this system was deemed more feasible. Remus et al. (2010) have created a German sensitivity dictionary named SentiWortschatz for the German language. For the purpose of creating a feeling dictionary, over 3500 German words were assigned positive and nega- tive values in the range of [-1, 1], using PosTags. Abdaoui et al. (2017) have created the FEEL: a French Expanded Emotion Lexicon polarity dictionary for French. Moreno-Sandoval et al. (2017) have created the Combined Spanish Lexicon polarity dictionary for Spanish.

Turkish Polarity Lexicon HisNet

In this study, we present a polarity dictionary to provide an extensive polarity dictionary for Turkish that dictionary-based sentiment analysis studies have been longing for. Our primary objective is to provide a more refined and extensive polarity dictionary than the previous SentiTurkNet. In doing so, we have resorted to a different network from the referenced study. We have identified approximately 76,825 synsets from Kenet, which then were manually labeled as positive, negative or neutral by three native speakers of Turkish. The first labelling process resulted in 3,100 positive, 10,191 negative and 63,534 neutral data, during which decisions were based on the meaning and connotation of each word.

Subsequently, a second labeling was further made on positive and negative words as strong or weak based on their degree of positivity or negativity. For instance, the word mükemmel (excellent) in Turkish has been marked three times. Thus, three different views were obtained for the value of this word. While selecting the appropriate label, the compatibility of the labels selected by the three labelers was also evaluated. To put it differently, if a positive word receives strong label from all three annotators, it is regarded as strong positive. If it receives two strong and one weak label, it is considered as very positive. If it is la- belled as strong once and as weak twice, it means it is just positive. Finally, if it receives weak label from all three annotators, it is considered as weak positive. The same is also true for the words labelled as negative.

Polarity Level # of Synsets
Strongly positive 1,038
Very positive 451
Positive 456
Weakly positive 1,234
Objective 63,534
Strongly negative 4,430
Very negative 1,465
Negative 1,238
Weakly negative 3,360

Video Lectures

For Developers

You can also see Cython, Java, C++, Swift, Js, or C# repository.

Requirements

Python

To check if you have a compatible version of Python installed, use the following command:

python -V

You can find the latest version of Python here.

Git

Install the latest version of Git.

Pip Install

pip3 install NlpToolkit-SentiNet

Download Code

In order to work on code, create a fork from GitHub page. Use Git for cloning the code to your local or below line for Ubuntu:

git clone <your-fork-git-link>

A directory called SentiNet will be created. Or you can use below link for exploring the code:

git clone https://github.com/starlangsoftware/TurkishSentiNet-Py.git

Open project with Pycharm IDE

Steps for opening the cloned project:

  • Start IDE
  • Select File | Open from main menu
  • Choose TurkishSentiNet-PY file
  • Select open as project option
  • Couple of seconds, dependencies will be downloaded.

Detailed Description

SentiNet

Duygu sözlüğünü yüklemek için

a = SentiNet()

Belirli bir alana ait duygu sözlüğünü yüklemek için

SentiNet(fileName: str)
a = SentiNet("dosya.txt")

Belirli bir synsete ait duygu synsetini elde etmek için

getSentiSynSet(self, _id: str) -> SentiSynSet

SentiSynSet

Bir SentiSynset elimizdeyken onun pozitif skorunu

getPositiveScore(self) -> float

negatif skorunu

getNegativeScore(self) -> float

polaritysini

getPolarity(self) -> PolarityType

Cite

@inproceedings{ozcelik21,
title={{H}is{N}et: {A} {P}olarity {L}exicon based on {W}ord{N}et for {E}motion {A}nalysis},
year={2021},
author={M. Ozcelik and B. N. Arican and O. Bakay and E. Sarmis and N. B. Bayazit and O. Ergelen and O. T. Y{\i}ld{\i}z},
booktitle={Proceedings of GWC 2021}
}

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

NlpToolkit-SentiNet-1.0.18.tar.gz (913.0 kB view hashes)

Uploaded source

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Fastly Fastly CDN Google Google Object Storage and Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page