This study investigates imbalanced class distributions in high-dimensional, sparse text corpora, where feature selection is often biased toward majority classes because dominant term frequencies overshadow minority evidence. Many conventional scoring functions fail to capture minority-class information adequately, thereby reducing categorisation effectiveness under skewed sampling. An additional challenge is selecting informative features while limiting the influence of ubiquitous, weakly discriminative terms that destabilise normalisation in sparse regimes. To address these issues, this paper proposes Chi-square Adaptive Weightage (CAW), an adaptive weight-scaling method that enhances minority recognition while maintaining stable, interpretable scoring under imbalance. CAW integrates three components: an adaptive weighting mechanism that compensates for uneven class frequencies during scoring, a normalised probability ratio that suppresses off-class and background evidence to retain class-specific informative terms, and a Chi-square component that statistically reinforces term-class dependence. Robustness is evaluated on two benchmark corpora with distinct characteristics, Reuters-21578 and Ohsumed, using four systematically sampled imbalance ratios (1:1, 2:1, 5:1, and 10:1) across three classifiers. Overall effectiveness is assessed using Macro-F1 and Micro-F1, while minority recognition is assessed using Minority Macro-F1 and mean AUPRC. Aggregated results show that CAW achieves the strongest Macro-F1 on Reuters across all ratios and remains top-tier on Micro-F1. Minority-focused evaluation under SVM provides direct evidence that CAW maintains minority influence as skew increases, ranking consistently in the top tier on Reuters under higher skew and remaining among the top methods on Ohsumed for Minority Macro-F1 and AUPRC. Overall, CAW offers a practical imbalance-aware feature selector for sparse text classification, with exceptionally reliable gains under linear decision boundaries.
TsaiCFChenKCLinWC. Feature selection and its combination with data over-sampling for multi-class imbalanced datasets. Appl Soft Comput2024; 153: 111267.
2.
ZhangCSodaPBiJet al.An empirical study on the joint impact of feature selection and data resampling on imbalance classification. Applied Intelligence2022; 53: 5449–5461.
3.
ReyCCGarcíaVSVilluendas-ReyY. Evolutionary feature selection for imbalanced data. 2023 Mexican international conference on computer science, ENC 2023, 2023. DOI: 10.1109/ENC60556.2023.10508674.
4.
LiuQCaiMLiQet al.Multi-label feature selection based on adaptive label enhancement and class-imbalance-aware fuzzy information entropy. Int J Approx Reason2025; 176: 109320.
5.
AhmadSRBakarAAYaakubMR. Ant colony optimization for text feature selection in sentiment analysis. Intell Data Anal2019; 23: 133–158.
6.
TarekegnANGiacobiniMMichalakK. A review of methods for imbalanced multi-label classification. Pattern Recognit2021; 118: 107965.
7.
ChakrabortySDeyL. Applications of ,ulti-objective, multi-label, and multi-class classifications. Singapore: Springer Nature Singapore, 2024135164. DOI: 10.1007/978-981-97-9622-9_6.
8.
BrownleeJ. Imbalanced classification with python: better metrics, balance skewed classes, cost-sensitive learning. San Juan, Puerto Rico: Machine Learning Mastery, 2020, https://books.google.com.my/books?id=jaXJDwAAQBAJ.
9.
SunLSunSet al.Feature selection using symmetric uncertainty and hybrid optimization for high-dimensional data. Int J Mach Learn Cybern2023; 14: 4339–4360.
10.
BhuiyanMSHRaianNALeonSIet al.Study of influence of dimension reduction of high dimensional datasets in classification problem. In: Proceedings of the 4th international conference on computing methodologies and communication, ICCMC 2020, 2020, pp.147–151. DOI: 10.1109/ICCMC48092.2020.ICCMC-00030.
11.
RayPReddySSBanerjeeT. Various dimension reduction techniques for high dimensional data analysis: a review. Artif Intell Rev2021; 54: 3473–3515.
12.
KowsariKMeimandiKJHeidarysafaMet al.Text classification algorithms: a survey. Information2020; 10: 150.
13.
ZhanZHShiLTanKCet al.A survey on evolutionary computation for complex continuous optimization. Artif Intell Rev2022; 55: 59–110.
14.
AbiodunEOAlabdulatifAAbiodunOIet al.A systematic review of emerging feature selection optimization methods for optimal text classification: the present state and prospective opportunities. Neural Comput Appl2021; 33: 15091–15118.
15.
FarghalyHMEl-HafeezTA. A high-quality feature selection method based on frequent and correlated items for text classification. Soft comput2023; 27: 11259–11274.
16.
AdelAOmarNAbdullahSet al.Co-operative binary bat optimizer with rough set reducts for text feature selection. Appl Sci2022; 12: 11296.
17.
AnsariGJShahJHFariasMCet al.An optimized feature selection technique in diversified natural scene text for classification using genetic algorithm. IEEE Access2021; 9: 54923–54937.
18.
AhmadISAbuARidzwanMet al.Beyond sentiment classification: a novel approach for utilizing social media data for business intelligence. Int J Adv Comput Sci Appl2020; 11: 437–441.
19.
YadavAKhannaMAnandD. A novel and efficient statistical and soft-computing intelligence integrated feature selection technique for human chronic diseases prediction. Multimed Tools Appl2025; 84: 41853–41896.
20.
SinghLKKhannaMGargHet al.A novel soft computing based efficient feature selection approach for timely identification of COVID-19 infection using chest computed tomography images: a human centered intelligent clinical decision support system. Multimed Tools Appl2024; 84: 13717–13785.
21.
AliMSJavedK. A novel inherent distinguishing feature selector for highly skewed text document classification. Arab J Sci Eng2020; 45: 10471–10491.
22.
ShrivastavaKKhannaM. CFA–ASA: an adaptive hybrid feature selection approach for thyroid disease prediction. SN Computer Science2025; 6: 781.
23.
ChandraA. Comparison of feature selection for imbalance text datasets. In: Proceedings of 2019 international conference on information management and t, ICIMTech 2019, 2019, pp.68–72. DOI: 10.1109/ICIMTECH.2019.8843773.
24.
ParlakBUysalAK. A novel filter feature selection method for text classification: extensive feature selector. J Inf Sci2023; 49: 59–78.
25.
CekikR. A new filter feature selection method for text classification. IEEE Access2024; 12: 139316–139335.
26.
KamalovFThabtahFLeungHH. Feature selection in imbalanced data. Ann Data Sci2023; 10: 1527–1541.
27.
ChenWYangKYuZet al.A survey on imbalanced learning: latest research, applications and future directions. Artif Intell Rev2024; 57: 137.
28.
HasaninTKhoshgoftaarTMLeevyJLet al.Severely imbalanced big data challenges: investigating data sampling approaches. J Big Data2019; 6: 107.
29.
SusanSKumarA. The balancing trick: optimized sampling of imbalanced datasets—a brief survey of the recent state of the art. Eng Rep2021; 3: e12298.
30.
KamalAHMZhuXPandyaA, et al.Feature selection for datasets with imbalanced class distributions. Int J Softw Eng Knowl Eng2010; 20: 113–137.
31.
RezvaniSWangX. A broad review on class imbalance learning techniques. Appl Soft Comput2023; 143: 110415.
32.
SuhaimiNSOthmanZYaakubR. Modified convolutional neural network filter gate for social media text classification. Int J Comput Sci Netw Secur2022; 22: 617.
33.
ParlakB. A novel feature ranking algorithm for text classification: brilliant probabilistic feature selector (BPFS). Comput Intell2023; 39: 900–926.
34.
CaiLJLvSShiKB. Application of an improved chi feature selection algorithm. Discrete Dyn Nat Soc2021; 2021: 1–8. DOI: 10.1155/2021/9963382.
35.
BeheraSKDashR. A novel feature selection technique for enhancing performance of unbalanced text classification problem. Intell Decis Technol2022; 16: 51–69.
36.
ChakrabortyJMajumderSMenziesT. Bias in machine learning software: why? how? what to do? In: Proceedings of the 29th ACM joint meeting on European software engineering conference and symposium on the foundations of software engineering. ACM, pp.429–440. DOI: 10.1145/3468264.3468537.
37.
LiuYLohHTSunA. Imbalanced text classification: a term weighting approach. Expert Syst Appl2009; 36: 690–701.
38.
TantisripreechaTSoonthornphisajN. A novel term weighting scheme for imbalanced text classification. Informatica2022; 46: 259–268.
39.
CuiYJiaMLinTYet al.Class-balanced loss based on effective number of samples. In: 2019 IEEE/CVF conference on computer vision and pattern recognition (CVPR). IEEE, pp. 9260–9269. DOI: 10.1109/CVPR.2019.00949.
40.
ArafIIdriAChairiI. Cost-sensitive learning for imbalanced medical data: a review. Artificial Intelligence Review 2024 57:42024; 57: 1–72.
41.
PesBLaiG. Cost-sensitive learning strategies for high-dimensional and imbalanced data: a comparative study. PeerJ Comput Sci2021; 7: e832.
42.
GulSKhanDMAldahmaniSet al.Margin weighted robust discriminant score for feature selection in imbalanced gene expression classification. PLoS One2025; 20: e0325147.
El-HajjWHajjH. An optimal approach for text feature selection. Comput Speech Lang2022; 74: 101364.
46.
LiuYJuSWangJet al.A new feature selection method for text classification based on independent feature space search. Math Probl Eng2020; 2020: 1–14.
47.
PalanivinayagamAEl-BayehCDamaŠevičiusR. Twenty years of machine-learning-based text classification: a systematic review. Algorithms2023; 16: 236. DOI: 10.3390/a16050236.
48.
KumarADabasVHoodaP. Text classification algorithms for mining unstructured data: a swot analysis. Int J Inf Technol2020; 12: 1159–1169.
49.
AlshalabiHATiunSOmarN. A comparative study of the ensemble and base classifiers performance in malay text categorization. Asia-Pac J Inf Technol Multimedia2017; 06: 53–64.
50.
AlyasiriOMCheahYNAbasiAKet al.Wrapper and hybrid feature selection methods using metaheuristic algorithms for English text classification: a systematic review. IEEE Access2022; 10: 39833–39852.
LeeMCHBraetJSpringaelJ. Performance metrics for multilabel emotion classification: comparing micro, macro, and weighted F1-scores. Appl Sci2024; 14: 9863.
53.
OpitzJ. A closer look at classification evaluation metrics and a critical reflection of common evaluation practice. Trans Assoc Comput Linguist2024; 12: 820–836.
54.
HancockJTKhoshgoftaarTJohnsonJ. Using Area Under the Precision Recall Curve to Assess the Effect of Random Undersampling in the Classification of Imbalanced Medicare Big Data. International Journal of Reliability, Quality and Safety Engineering2024; 31: 2350039. DOI: 10.1142/S0218539323500390.
55.
LeevyJLKhoshgoftaarTMHancockJ. Evaluating performance metrics for credit card fraud classification. In: Proceedings - international conference on tools with artificial intelligence, ICTAI 2022, 2022 October, pp.1336–1341. DOI: 10.1109/ICTAI56018.2022.00202.
56.
HancockJTKhoshgoftaarTMJohnsonJM. Evaluating classifier performance with highly imbalanced big data. J Big Data2023; 10: 42.
