Scholarlink

<p>Tomato (<i>Solanum lycopersicum</i> L.) is the second most widely grown vegetable crop globally. Enhancing yield is still a challenge due to the complexity of yield-attributing characters. Fourteen tomato genotypes were tested across four seasons (2016–2020) under open field conditions for 18 quantitative characters. Three prediction models viz., multilayer perceptron (MLP), multiple linear regression (MLR), and Classification and Regression Tree (decision tree, CART algorithm) were used to identify the most important characters influencing yield per plant (YPP). The MLP identified fruits/plant (FPP) and ascorbic acid (AC) as important yield determinants with low training and testing errors, reflecting high prediction quality. The MLR analysis also validated significant positive influence of FPP and fruit weight (FW) towards yield/plant with an adjusted R<sup>2</sup> value of 0.95. The contribution of fruits/plant towards yield/plant alone was 74%. Titrable acidity (TA), lycopene content (LC), total sugar content (TS) and polar diameter (PD) showed negative influence, indicating trade-offs between yield and some quality characters. The Decision Tree model determined threshold-based character splits, validating FPP, flowers/cluster (FC), and FW as the most important factors contributing to YPP. This research illustrated the promise of AI-based phenotyping in trait prioritization and pre-emptive breeding, providing a solid, data-driven platform to expedite genetic gain in tomato.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Integrating artificial intelligence (AI) based models for yield prediction in tomato (Solanum lycopersicum L.)

  • 期刊论文

  • 发表日期：

    2026年5月27日

查看全文
  • Soham Hazra,
  • Suvojit Bose,
  • Subhadwip Ghorai,
  • Ankur Mukhopadhyay,
  • Avishek Chatterjee,
  • Poulomi Sen,
  • Rajdeep Mohanta,
  • Pranab Hazra,
  • Sourav Roy

摘要

Tomato (Solanum lycopersicum L.) is the second most widely grown vegetable crop globally. Enhancing yield is still a challenge due to the complexity of yield-attributing characters. Fourteen tomato genotypes were tested across four seasons (2016–2020) under open field conditions for 18 quantitative characters. Three prediction models viz., multilayer perceptron (MLP), multiple linear regression (MLR), and Classification and Regression Tree (decision tree, CART algorithm) were used to identify the most important characters influencing yield per plant (YPP). The MLP identified fruits/plant (FPP) and ascorbic acid (AC) as important yield determinants with low training and testing errors, reflecting high prediction quality. The MLR analysis also validated significant positive influence of FPP and fruit weight (FW) towards yield/plant with an adjusted R2 value of 0.95. The contribution of fruits/plant towards yield/plant alone was 74%. Titrable acidity (TA), lycopene content (LC), total sugar content (TS) and polar diameter (PD) showed negative influence, indicating trade-offs between yield and some quality characters. The Decision Tree model determined threshold-based character splits, validating FPP, flowers/cluster (FC), and FW as the most important factors contributing to YPP. This research illustrated the promise of AI-based phenotyping in trait prioritization and pre-emptive breeding, providing a solid, data-driven platform to expedite genetic gain in tomato.