So far, no method (that I know of) has shown itself to have a "remarkable" amount of statistical forecasting accuracy for intra-state conflict. That said, I suggest that you take a look at the work of political demographers at PRIO in Norway who have tried to improve forecasts by including a measure of age-structure (see papers by Henrik Urdal, Havard Nygard, and others).  The advantage to using age structure (young-adult proportion (15-24 yrs of age) divided by the adult proportion of the total population (15+); or the median age) is that future risks can be calculated from the UN Population Division's future projections (or, in the case of median age, obtained directly from projection data, see UN Population Division, "Population Prospects, the 2015 Revision", online).  

From my own work, I find that: (a.) countries that have been in intra-state conflict at least one year during the past 5 years (according to the Uppsala U./PRIO database) are the most likely to experience the initiation of conflict (a somewhat trivial finding, since much of it is a re-ignition of former/ongoing conflicts); (b.) among countries without a conflict in the past 5 years, those with youthful age structures -- a median age less than 26 years -- have an elevated likelihood of experiencing the INITIATION of a new low-intensity or high-intensity intra-state conflict within the next decade. The youngest states, below a median age of 20, are at highest risk.  (c.) Among states older than a median age of 26, very few experience the initiation of a intra-state conflict.  (d.) Among those that do experience the onset of an intra-state conflict, youthful minorities are principally the "focus" of those conflicts (with only a few exceptions, e.g., the Ukraine).  

Note that the UN Population Division's median age estimate/projections for the Gulf Cooperation Council states (Saudi Arabia, Bahrain, UAE, Oman, Kuwait, Qatar) are not used in this analysis.  The UN includes temporary migrants, making those populations look much older (and less volatile) than they really are.  I adjust these values based on estimates/projections of citizen residents that I obtained from other sources.   Hopefully, I will have these data online this year.  

(in my opinion) Age structure provides a good indication (NOT an excellent indication) of the potential volatility of states. It identifies a group of states with either (1.) a high risk of the initiation of an intra-state conflict or (2.) the presence of a persistent autocracy (Freedom House "Not Free").  Can it be used to predict conflict? -- perhaps over a 10 year time frame.  

A criticism of this method is that it is overly simple.  I argue that it is a measure that integrates the most important structural measures: fertility, educational attainment, and household income, all of which are related.  Age structural "maturity" is a product of fertility decline, and affects (and is affected by) educational attainment (see papers by Ronald Lee and Andrew Mason). Educational attainment affects (and is affected by) household income.  The links between fertility and household income are significantly weaker, but probably exist (savings, and women's employment).  

 I've attached a short paper.  Work by Urdal, Nygard and others at PRIO are on Researchgate and Academia.edu. I hope this is helpful,

Regards, Richard Cincotta

Article Population Age Structure and Its Relation to Civil Conflict:...

Thank you so much for your answer