Application of topic modelling to Integrated Water Resource Assessment domain
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Integrated Water Resource Assessment (IWRA) is a multidisciplinary analysis of cause-effect interactions of environmental, social, and economic processes which play an important role in an investigated scenario or undertaking related to a water resource. The goal of this analysis is to generate new knowledge in support of decision or policy making. The concept of IWRA covering natural processes in a water resource with at least one of the areas of economic or social aspects of modern society had been articulated almost a half a century ago. Nowadays, it became mature to the extent of supporting environmental sustainability and promoting UN Sustainable Development Goals through offering well-accepted approaches, frameworks, simulation models and computational techniques upholding the assessment. Nevertheless, there is steady interest to the issues of IWRA among researchers and practitioners because new technologies open opportunities for advanced computational techniques and comprehensive analysis. This study presents exploratory analysis of the corpus of scientific publications using text mining techniques with the aim to identify salient topics and potential gaps in the IWRA research area. The analysis was conducted based on the topic modelling approach. Topic modelling is a form of text mining that allows to find a representation of information from a collection of documents called corpus. Any text document can be viewed as a collection of several themes which are present in the document and reflect the document contents in a meaningful to its readers way. A theme or a topic is represented via an array of words that have a high tendency of co-occurrence when a particular theme underlying a document is being discussed. The most salient characteristic of topic models is that they automate the process of extracting these underlying (latent) themes in large corpora of texts without any human intervention excluding text pre-processing. Given that a topic model operates with a fixed vocabulary, domain specific analysis is expected to be more informative. Therefore, careful selection of documents included into a corpus is required. Application of topic modelling to multidisciplinary areas such as IWRA carries more importance because it helps to automate the process of extraction of salient topics relevant to a document and categorize the documents into themes for targeted analysis and knowledge extraction. The corpus of abstracts of 89726 papers published from 1970 to 2020 in peer-reviewed journals representing leading outlets in the areas of water resources and integrated environmental assessment was assembled. It was analysed using basic bibliometric statistics. After that, the corpus was pre-processed following conventional topic modelling framework and fed into LDA mallet algorithm to identify salient topics. Hyperparameters of the selected topic modelling algorithm were identified based on exploratory computations and evaluation of several topic models performance using a coherence score and qualitative evaluation of the identified topics. The model producing 20 topics was considered satisficing and used as a basis for the qualitative analysis of clusters of words forming topics. The analysis revealed two categories of latent topics presented in the corpus: methodological and environmental. The latter describes various aspects of utilization, protection, and restoration of a natural water resources. No theme reflecting assessment of socio-economic processes was uncovered despite the fact, that these processes play critical role in the environmental state of a water resource.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it