基于多源机器学习的脱贫方式智能推荐研究

Intelligent Recommendation for Poverty Alleviation Based on Multi-Source Machine Learning

由于贫困户自主选择帮扶政策和人工推荐帮扶政策缺乏客观性和公正性,难以推荐合理的帮扶政策。因此,利用多种机器学习分类算法和多源集成实现贫困户帮扶政策的智能推荐。结果表明智能算法给出的扶贫方式在帮扶政策的推荐上具有近90%的准确率,有助于政府扶贫工作人员帮助扶贫对象进行帮扶政策的选择,达到精准扶贫的目的。

Due to the lack of objectivity and impartiality of poverty-stricken households'self-selection of support policies and manual recommendation assistance policies,it is difficult to recommend reasonable assistance policies.Therefore,this paper uses a variety of machine learning classification algorithms and multi-source integration to achieve intelligent recommendation of poor households'support policies.The results show that the poverty alleviation methods given by intelligent algorithms have an accuracy rate of about 90%in the recommendation of support policies which could help the government's poverty alleviation workers give the poor reasonable advice about assistance policies and achieve the goal of precision poverty alleviation.