Development of a Semantic Text Classification Mobile Application Using TensorFlow Lite and Firebase ML Kit
Bibliographic record
Abstract
The development of neural networks in the current industrial era 4.0 should help various work fields, one of which is the scientific literature.The problem that often occurs is that scientific papers still use manual sorting of themes/semantics.The purpose of this research is to build a semantic text classification application that can allow users to sort by theme/semantics by using a neural network model, Recurrent Neural Network (RNN) embedded in a smartphone.The development of this application uses the waterfall method in which there are analysis and system design.The application implements the text recognition feature of the Firebase ML Kit.It is developed using a general machine learning cycle method or approach consisting of data identification, data preparation, algorithm selection, model training, model evaluation and model deployment.The model was built using abstract data from scientific papers from the State University of Padang Library.The total data obtained 84 training data and 21 test data using a ratio of 80:20 percent to perform the validation test.The neural network model uses the AverageWordVec specification provided by TensorFlow Lite Model Maker with three classification outputs.The model validation test reached 0.7619 accuracy values with 0.7782 loss values.The model is executed using the TensorFlow Lite interpreter embedded in the application.The application results fulfill the overall system functional requirements analysis.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.014 | 0.006 |
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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".